BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY VOL. VI 1959-1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM (NATURAL HISTORY) LONDON: 1960 DATES OF PUBLICATION OF THE PARTS No. i. 23 October 1959 No. 2. 19 February 1960 No. 3. n March 1960 No. 4. ii March 1960 No. 5. ii March 1960 No. 6. 12 April 1960 No. 7. 22 July 1960 PRINTED IN GREAT BRITAIN AT THE BARTHOLOMEW PRESS DORKING BY ADLARD AND SON, LTD. CONTENTS ZOOLOGY VOLUME 6 PAGE No. i. Taxonomy of the genus Columba. By D. GOODWIN i No. 2. A revision of the Platyseiinae (Mesostigmata: Aceosejidae) based on material in the collections of the British Museum (Natural History) . By G. OWEN EVANS and K. H. HYATT 25 No. 3. British slugs (Pulmonata: Testacellidae, Arionidae, Limacidae). By H. E. QUICK (Pis. 1-2) 103 No. 4. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae) Part IV. By P. H. GREENWOOD 227 No. 5. On a Stygiomysis from the West Indies, with a note on Spelaeogriphus (Crustacea, Peracarida). By ISABELLA GORDON (Pis. 3-4) 283 No. 6. Contributions toward a classification of modern Isospondylous fishes. By WILLIAM A. GOSLINE 325 No. 7. The ecology and taxonomy of some Angola birds (Based on a collection made in 1957). By B. P. HALL (Pis. 5-6) 367 Index to Volume 6 455 TAXONOMY OF THE GENUS COLUMBA D. GOODWIN BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. i LONDON: 1959 TAXONOMY OF THE GENUS COLUMBA BY D. GOODWIN Pp. 1-23 ; 5 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. i LONDON : 1959 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. i of the Zoological series. Trustees of the British Museum, 1959 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued October, 1959 Price Nine Shillings TAXONOMY OF THE GENUS COLUMBA By D. GOODWIN CONTENTS Page INTRODUCTION ........... i COLOUR AND COLOUR-PATTERN . . . . . . . i ADAPTIVE RADIATION AND RELATIONSHIPS ...... 2 THE SPECIES-GROUPS AND THEIR COMPONENT SPECIES .... 4 NOTES ON SPECIES .......... 7 SUMMARY ............ 22 REFERENCES ........... 22 SYNOPSIS Previous reviews of the genus Columba have grouped together species which lack display plumage on the neck or whose ornamentation is superficially similar. This is believed to have resulted in some false conclusions because such display plumage has been independently lost or become obsolescent in several species. When this is recognized and greater importance attached to basic colour-patterns and geographical distribution a truer picture of relationships within the genus can be arrived at. INTRODUCTION THE genus Columba includes a large number of medium to large-sized pigeons, widely distributed throughout the world. Peters (1937), for example, lists fifty-two species and does not divide them into sub-genera. The rather inconveniently large size of the genus has often been criticized and the affinities of its members questioned. Ridgway (1916) separated the relatively few northern and central American species into six genera, some of them monotypic. The most important recent revision of the genus is that of Boetticher (1954), who divides it into several sub-genera and corrects some false conclusions of a previous and less comprehensive revision by Miculicz- Radecki (1949). When re-arranging the very extensive collection of specimens of Columba in the British Museum (Natural History) it was found impossible to agree with all of Boetticher's conclusions. The object of this paper, therefore, is a revision of the genus based on examination of the skins in the National Collection and correlated, as far as has been possible, with what is known of the ecology and behaviour of the species of which it is comprised. Emphasis is placed on discussion of those points where my conclusions differ from those of Boetticher. I am indebted to Dr. Dean Amadon of the American Museum of Natural History for the loan of specimens of Columba albinucha and C. pallidiceps, to Mr. J. D. Macdonald for help and con- structive criticism and to Dr. G. Bodenstein, Professor W. F. Hollander and Mr. R. W. Sims for information and discussion on various points. COLOUR AND COLOUR-PATTERN Within Columba plumage colour ranges from entirely black, with green and purple iridescence, to predominantly pale grey or purplish-chestnut. It is evident that, ZOOL. 6, I. I TAXONOMY OF THE GENUS COLUMBA as in Ptilinopus (Cain, 1954) colour-pattern is usually a better guide to relationship than actual colour. One finds striking differences in the colour of the plumage and/ or soft parts of the head and neck (the focal area of both sexual attraction and aggres- sive attack in pigeons) between pairs of sympatric and very closely related species as C. squamosa and C. leucocephala or C. rupestris and C. leuconota (see Text-fig, i). KEY TO SHADING Dull purple or purplish-brown Iridescent green &XX3 Iridescent purple or wine-red Black or greyish-black, glossed with green in fallidiccfs White or silver-grey y. i { { i\ Display plumage with dark feather- ) ) ) ) ) edges giving a scale-like appearance FIG. i. Diagrammatic sketches to show differences in head and neck colours between three pairs of sympatric and closely related species. From top left : pallidiceps and vitiensis, squamosa and leucocephala, rupestris and leuconota. Such differences almost certainly function as isolating mechanisms to inhibit inter- specific pairing. Signal markings on wings and tail and specific differences in voice and display-flight serve the same purpose at " longer range ". As with other groups of birds the greatest diversity of signal markings, display plumage and voice is found where there are many related and at least potentially sympatric species and, conversely, a tendency to lose such characters and acquire a dull uniformly coloured plumage in isolated island forms. Lack of signal markings or display plumage in Columba seems never to be a primitive character. Examination of the colour- pattern and other characters of the species concerned shows convincingly that their common lack of ornamentation is due to convergence and not to close affinity. Sexual dimorphism in Columba is relatively slight but in most species it is sufficient to enable skins or living birds to be correctly sexed on plumage colour alone. In those species which are predominantly grey males are a purer, bluer grey than females. Purple, chestnut or mauve-pink in males is often duller, absent, or replaced by dull brown in females. Iridescent display plumage (when present) is usually more brilliant in males. Where adults show such differences of plumage similar and homo- logous sexual differences are found in the juvenile plumage. ADAPTIVE RADIATION AND RELATIONSHIPS Adaptive radiation can be misleading at generic level, since once two forms have evolved marked differences of size and structure they are unlikely to be considered TAXONOMY OF THE GENUS COLUMBA 3 con-generic. For example the many Australian bronze-wing pigeons are obviously very close relatives in spite of the taxonomically convenient differences in size and proportions which they have developed in the course of adaptation to different ecological niches. Within Columba, as here recognized, the greatest amount of adaptive radiation is found among the Palearctic and Ethiopian species. Similar, but less marked differences are shown among the American species, but here maculosa, which otherwise closely parallels the old-world lima, has not emancipated itself from the use of trees for resting and breeding. The question arises whether the New World representatives of Columba should be generically separated from the Old World forms. Cumley & Irwin (1944) concluded from their work on the antigenic inter-relationships of the blood cells of guinea, livia, palumbus, janthina, fasciata, maculosa, picazuro, cayennensis, flavirostris and leucocephala that each of the six American species was more closely related to the other five than to any of the four Eurasian species. Verheyen (1957) considers that skeletal differences warrant the separation of Old and New World forms of Columba, both of which groups he further divides into several genera, some of them monotypic. I am not competent to criticize Verheyen's opinions in so far as they are based on osteology, but his decisions placing guinea in the monotypic genus Dialiptila and generically separating janthina, vitiensis and norfolciensis are so contrary to the evidence from external taxonomic characters and behaviour-patterns that I hesitate to follow his conclusions without further evidence. If the New World species are to be generically separated this could be done only on osteological and biochemical characters. Possibly studies currently in progress in America will show that there are constant and significant behavioural differences also and it will be desirable to separate the American from the Old World species. I think, however, that at present it would be premature to do so. It seems to me best to retain the large genus Columba as used by Peters (1937) and Hellmayr & Conover (1942) but to give Oenoenas and Turturoena sub-generic rank within it. I also think that Nesoenas should be given only sub-generic rank within Columba. The remaining forms can be relegated to five species-groups. Cain (1954), in his revision of the pigeons in the genus Ptiliuopus, used the category of the species-group for " an informal natural group of closely related species below the rank of a subgenus, the recognition of which involves no complication of the scientific names of the species contained in it." And that of sub-group for more closely related forms within the species-group. I have followed Cain in the use of these categories which are equally useful for defining relationships of the pigeons hi Columba. The hierarchy of categories is therefore : Genus Sub-genus Species-group Sub-groups (of species group) Superspecies Species Subspecies ZOOL. 6, I. r 4 TAXONOMY OF THE GENUS COLUMBA There are, naturally, some species whose characters show them to be inter- mediate between two sub-groups and it has been necessary to allot them somewhat arbitrarily to that with which they seem to have most in common or else to place them in a mono-specific sub-group. All such cases are discussed below. In the following list members of the same superspecies are bracketed together. Sub- species have not been listed but they are discussed under the species headings where they are relevant to specific problems. THE SPECIES-GROUPS AND THEIR COMPONENT SPECIES (i) Oenas Species-group Smallish (oliviae) to large (trocaz) sized columbids inhabiting the Palearctic and Ethiopian regions, one species (livia) also occurring in India, and throughout much of the world in a feral state as a result of human introduction. Most species in this group have well developed display plumage on the neck and patterned tails. Sexual dimorphism slight. (a) Palumbus Sub-group Tail rather long with conspicuous pale central band ; breast mauve-pink or salmon-pink. C. palumbus Linnaeus, 1758 C. trocaz Heineken, 1829 C. bollii Godman, 1872 C. unicincta Cassin, 1859 (1860) (b) Oenas Sub-group Glossless mauve-pink on head and/or breast, black signal markings on inner wing-coverts (obsolescent in oliviae). C. oenas Linnaeus, 1758 C. eversmanni Bonaparte, 1856 C. oliviae Stephenson Clarke, 1918 (c) Livia Sub-group No glossless mauve-pink colour on breast or head. Display plumage on front as well as sides of neck (absent in leuconota). C. livia Gmelin, 1789 C. rupestris Pallas, 1811 C. leuconota Vigors, 1831 C. guinea Linnaeus, 1758 C. albitorques Riippell 1837 TAXONOMY OF THE GENUS COLUMBA 5 (2) Junioniae Species-group Large columbid inhabiting the western Canary Islands of Palma and Gomera. Mantle and wings dark brown and immaculate ; tail with broad pale terminal band. Entire neck and breast iridescent but no well differentiated display plumage. One species C. junioniae Hartert, 1916. (3) Elphinstonii Species-group Smallish medium to large-sized columbids inhabiting the Indo-malayan and Ethi- opian regions. All except one with unpatterned tails. No signal markings on wings. Most with display plumage on sides and back of neck. (a) Arquatrix Sub-group Display plumage on hind neck lanceolate. Wing-coverts and /or underparts spotted with white (obsolescent spots in polleni}. {C. arquatrix Temmink, 1809 C. pollenii Schlegel, 1866 C. hodgsonii Vigors, 1832 C. albinucha Sassi, 1911 (b) Elphinstonii Sub-group Display plumage on hind neck of black feathers tipped white or buff. Iridescent green or purple sheen on neck and mantle. Unpatterned tails, no signal markings on wings. One superspecies. {C. pulchricollis Blyth, 1845 (1846) C. elphinstonii (Sykes, 1832) C. torringtoni Bonaparte, 1854 (c) Punicea Sub-group No well differentiated display plumage on neck ; wing coverts and body plumage immaculate chestnut or silver contrasting with black primaries. C. punicea Blyth, 1842 C. argentina Bonaparte, 1855 (d) Palumboides Sub-group Head and neck silver-grey ; wings and back dark grey with iridescent feather edges. One species. C. palumboides (Hume, 1873) (4) Janthina Species-group Large columbids inhabiting the Pacific regions from Japan to Australia, chiefly on islands. Plumage largely or entirely blackish and highly iridescent. No signal markings on wings ; dark unpatterned tails. Not divisible into sub-groups. 6 TAXONOMY OF THE GENUS COLUMBA {C. janthina Temminck, 1830 C. vitiensis Quoy & Gaimard, 1830 C. norfolciensis Latham, 1801 C. versicolor Kittlitz C. jouyi Stejneger, 1887 C. pallidiceps (Ramsay, 1877) (5) Picazuro Species-group Small (rufina) to medium-large (maculosa) columbids inhabiting the nearctic and neotropic regions. Most with display plumage on back and sides of neck and patterned tails. (a) Leucocephala Sub-group Patch of velvet-like feathers on nape ; well-defined iridescent display plumage on back and sides of neck. Plumage largely dark slate-grey ; wings and tail unmarked. C. leucocephala Linnaeus, 1758 C. squamosa Bonnaterre, 1792 (b) Speciosa Sub-group Feathers of neck and breast white and golden with dark, iridescent edges ; back, rump and wing-coverts purplish-chestnut (male) or dark brown (female). Tail blackish and unmarked. One species only. C. speciosa Gmelin, 1789 (c) Picazuro Sub-group Outer wing-coverts edged white ; mantle and wing-coverts brown, tail with ill- defined terminal dark band. C. picazuro Temminck, 1813 C. corensis Jacquin, 1784 C. maculosa Temminck, 1813 (d) Fasciata Sub-group Parti-coloured tail with broad, pale terminal band. Iridescent green display plumage on hind-neck. One superspecies. {C. fasciata Say, 1823 C. araucana Lesson, 1827 C. caribaea Jacquin, 1784 (e) Rufina Sub-group Small size. Crown and nape iridescent ; ill-defined pale terminal band on tail; plumage mainly vinous-purple. C. rufina Temminck, 1810 TAXONOMY OF THE GENUS COLUMBA 7 (f) Flavirostris Sub-group No iridescence on head or neck ; plumage largely vinous-purple ; tail dark grey without markings. One superspecies. {C '. flavirostris Wagler, 1831 C. oenops Salvin, 1895 C. inornata Vigors, 1827 Sub-genus Oenoenas Small columbids of central and southern America with dark brown and greyish- purple or reddish-purple plumage ; small bills, strongly rounded tails, obsolescent display plumage on hind neck and no markings on wings or tail. Consists of only four species. It seems purposeless to sub-divide further this small and very homo- geneous assemblage into sub-groups, which could only be done, if at all, on the shorter tails of two of them, nigrirostris and goodsoni. C. subvinacea (Lawrence, 1868) C. plumbea Vieillot, 1818 C. nigrirostris Sclater, 1859 (1860) C. goodsoni Hartert, 1902 Sub-genus Turturoena Small columbids inhabiting the Ethiopian region. Bright metallic iridescence on hind-neck. Marked sexual dimorphism except in the island form malherbii. One superspecies. {C. delagorguei Delagorgue, 1847 C. iriditorques Cassin, 1856 C. malherbii Verreaux & Verreaux, 1851 Sub-genus Nesoenas Large-sized Mauritian columbid with rounded wings (first primary about equal in length to sixth) and rufous tail. One species. C. mayeri Prevost, 1843 NOTES ON SPECIES C. palumbus The Wood Pigeon is considered by Boetticher (1954) to be most closely related to torringtoni, elphinstonii , trocaz and junioniae and to be much nearer to fasciata and allied American species than to oenas or unicincta. He grouped these, and some other pigeons, close together because they all show a high degree of ornamentation on the neck feathers, similar relative lengths of wing and tail and arboreal habits. Certainly the conspicuous white or cream-coloured patches on its neck give palumbus a superficial resemblance to such species as fasciata which also have con- spicuous white areas in the display plumage of the neck. In its colour-pattern and TAXONOMY OF THE GENUS COLUMBA TAXONOMY OF THE GENUS COLUMBA g even in the individual colours, however, it agrees more closely with unicincta and oenas. The difference in coloration of the display plumage on the necks of oenas and palumbus, due to the white patches of the latter, can be interpreted as an iso- lating mechanism (see Sibley, 1957) as can the difference in their advertising calls. The bowing display of palumbus is identical with that of oenas and its pre-copulatory behaviour is very similar to that of oenas, guinea and lima (see Goodwin, 19560) and differs from that oi fasciata as described by Whitman (1919). It is true that in some points of structure and deportment palumbus bears more resemblance to fasciata than to oenas but this may be attributed to the fact that palumbus and fasciata are largely arboreal and obtain much of their food from the tips of slender branches whereas oenas is almost entirely a ground feeder. Geographical distribution also seems to favour the linkage of palumbus and oenas. In my opinion, and excepting obvious derivatives on the Atlantic islands (see trocaz and bollii], palumbus has closest affinities with unicincta and is nearer to oenas than it is to pulchricollis or fasciata or to their close allies. At the same time its undoubted resemblance to these latter emphasize the difficulties of generically separating the Eurasian and American groups of Columba. C. trocaz The Madeiran Pigeon appears to be a fairly recent derivative of palumbus. In its voice and bowing display it closely resembles that species and in appearance it differs only in its slightly darker coloration, in lacking the white signal markings on the wings and in its slightly longer legs and more sloping forehead. As, however, it would seem to have diverged sufficiently not to have interbred with later Wood Pigeon invaders, now distinguished racially as C. palumbus maderensis, it has to be accorded specific rank. C. bollii Bolle's Pigeon is considered by Peters (1937) and others to be conspecific with trocaz. Both appear to be derivatives of palumbus but evidence suggests that they did not share a common ancestry subsequent to their differentiation from the parent stock. Although both have lost the white wing patch of palumbus the white neck patch has modified in different directions in each. In trocaz it has become a dull silvery grey and increased in area whereas in bollii it has become a small glossy brown mark. Since trocaz cannot now be considered conspecific with palumbus it seems better to give bollii specific status also. C. unicincta Boetticher considers the Afep Pigeon to be a relatively primitive species, most closely related to maculosa, because it lacks display plumage on the neck. It seems to me much more likely that loss of display plumage has occurred at a relatively late stage in the evolution of both species and does not indicate any close phylo- genetic relationship between them. In its general colour pattern unicincta agrees fairly closely with palumbus and its tail markings are identical with those of that io TAXONOMY OF THE GENUS COLUMBA species. Two features in which unicincta differs from palumbus are in its pale-edged wing-coverts and in its grey feet and bill. Since, however, these features, like the lack of iridescence on the neck, are similar to those in juveniles of palumbus I do not think they diminish the likelihood of close phylogenetic relationship of unicincta and palumbus, which their other features, including the little that has been recorded about the behaviour and ecology of unicincta, seem to indicate. C. oenas The Stock Dove is usually considered to be much more closely related to lima than to palumbus but in fact it forms a connecting link between the two in its taxo- nomic characters as well as in its ecology and behaviour (Goodwin, ig56a). Thus it agrees with palumbus in the distribution of the display plumage on its neck, in its glossless mauve-pink breast and in the form of its bowing display, but its wing and tail markings and its proportions are closer to lima. In view of the intermediate position of this, the type species of Columba, generic separation of " wood pigeons " and " rock pigeons " becomes untenable. C. eversmanni Eversmann's Stock Dove is extremely close to oenas in appearance, differing only in its smaller size, yellow irides and in having the head as well as the breast mauve- pink, or with a strong mauve-pink tinge, and in sometimes having a white instead of a grey rump. I concur with Peters in thinking that the evidence at present available warrants their both being considered full species. Stuart Baker (1928) put eversmanni as a race of oenas stating that " intermediate forms are not infrequent ". He gave no evidence for this other than mentioning some grey-rumped eversmanni having been obtained in Afghanistan. Except in this one feature grey-rumped individuals do not resemble oenas any more closely than do white-rumped ones. Stuart Baker's further remarks about the affinities of the two show that he knew little about the behaviour and ecology of oenas. The breeding range of eversmanni as indicated by Dementiev et al. (1951) has a considerable overlap with that of oenas. That they retain their respective characteristics where they are, apparently, sympatric, is evidence of their specific status. C. oliviae The Somaliland Pigeon can be regarded as forming a superspecies with eversmanni. The differences between them can be explained by the adaptation of oliviae to its torrid desert habitat. As well as acquiring a paler and browner coloration it has lost the black wing bars, probably quite recently, as many specimens show a few blackish flecks where these are in eversmanni. The loss of these signal markings may have resulted from a need for more perfect camouflage or simply to lack of selection for them subsequent to isolation from other closely related species. Unfortunately nothing appears to have been recorded of the behaviour or breeding habits of this species. In spite of its alternative vernacular name of " Somali Rock Pigeon ", oliviae is no more closely related to livia than are eversmanni or oenas. TAXONOMY OF THE GENUS COLUMBA n C. al orques The White-collared Pigeon has a white neck ring like that oifasciata and acuminate neck feathers like arquatrix. But as it resembles these species in no other respects it is highly unlikely that these likenesses indicate close relationship. Its tail pattern and the black markings on its wings seem to indicate relationship to oenas and lima. Taibell (1954) describes a bowing display like that of oenas. The lack of strong contrast between the black wing markings and their dark grey background, the slight degree of iridescence of the neck feathers and the partially concealed white wing patch may indicate that albitorques is at a stage of evolution where it is beginning to lose some of its display and signal markings, presumably as a result of being now ecologically isolated from related species. On the whole the appearance and taxonomic characters of this pigeon suggest a rather closer relationship with lima and guinea than with oenas and I have, accordingly, placed it in the lima sub-group. C. guinea It is fairly certain that the Speckled Pigeon is a close relative of lima (see Goodwin, 1956 and 19560) in spite of their differences of colour. They resemble each other in having display plumage on the front as well as on sides and back of the neck and in having the feathers of this display plumage bifurcated ; there are also many close resemblances of voice and behaviour. On the evidence of the blood antigens Cumley & Irwin showed that biochemically guinea is closer to palumbus than it is to lima or oenas. This is supported by two points of behaviour its calls and postures immediately after copulation and its methods of fighting. Thus what is known of this species emphasizes the essential homogeneity of the oenas species-group in spite of the rather considerable, if somewhat superficial, taxonomic and behavioural variations within it. I think, however, that in sub-dividing this group guinea should be put in the same sub-group as lima. C. leuconota The Snow Pigeon's plumage and general habits suggest a close relationship with lima and rupestris. The great differences in colour of the head and neck between leuconota and rupestris is almost certainly to be considered as an isolating mechanism and of little importance as an indicator of phylogenetic disparity. The same may be true of the apparently considerable difference between the voice and bowing display of leuconota (Newman, 1911) and that of lima. The black and white pattern on the tail of leuconota appears different from that of livia and other allied species but is essentially similar, the differences being due to a reduction of the black terminal band and an extension of the black basal area on the outer rectrices. C. livia The Rock Pigeon gives an impression of standing far apart from palumbus in appearance and behaviour. Boetticher (1954) indeed puts them as the terminal twigs on the two mutually diverging main branches of his family tree of Columba. The differences of form, behaviour and coloration shown by livia are largely due 12 TAXONOMY OF THE GENUS COLUMBA to its adaptations to nesting and sheltering in caves and rock fissures and feeding in open tree-less country. In fact I think that lima and palumbus are fundamentally much alike and that they represent the extremes in a single rather large species-group. Although lima now occurs in some moderately wooded regions it almost certainly evolved in a treeless or almost treeless environment. Its points of resemblance to rupestris, guinea and oenas are discussed under those species. C. rupestris The Blue Hill Pigeon is usually accorded specific rank although Austin (1948) put it as a race of livia but without giving reasons for this decision. Over most of their range the two are allopatric, rupestris replacing livia in the higher and colder parts of central and eastern Asia. The differences between them are certainly such as suggest racial rather than specific status. Rupestris has a proportionately smaller bill and a reduced amount of melanin in the plumage resulting in a white instead of a grey central bar in the tail, smaller black wing bars, paler grey ground colour and less intensely iridescent display plumage ; all differences such as might be expected in a race inhabiting a colder climate. With the possible exception of the white tail bar none of these cause as great difference in appearance as exists, for example, between C. livia gymnocyclus and C. /. dakhlae or C. L livia and C. I. intermedia. In its habits and ecology (Schafer, 1938) rupestris apparently differs in no essentials from livia. These facts suggest that the two are conspecific. On the other hand rupestris is said to associate with livia in parts of northern India (Stuart Baker, 1913) and it is potentially in contact with feral specimens of livia in many parts of its range. La Touche (1934) shot a specimen of C. livia intermedia (sic) out of a flock of rupestris in eastern China. The voice of rupestris, as described by Salim Ali (1949), would appear to differ much from that of livia. The evidence is thus conflicting. La Touche considers that the occurrence of mixed flocks suggest that rupestris and livia inter- breed but this might equally well be evidence to the contrary. If, as is likely, the term " mixed flock " has been applied to aggregations of rupestris and (? feral) livia at feeding grounds then such associations are of no significance either way. There are no specimens from among the many from North India and the Himalayas in the National Collection that appear to be hybrids or intergrades between the two. It must be admitted, however, that such birds might be difficult to detect, unless they were intermediate in coloration of the tail. Until the matter can be investigated more conclusively it seems best to accept the majority opinion that rupestris is a full species. Since the evidence suggests strongly that, at least as far as genuinely wild individuals are concerned, livia and rupestris are allopatric when breeding they should, I think, be considered members of a superspecies. C. junioniae The coloration and size of the Laurel Pigeon suggest relationship to palumbus. On the other hand the colour-pattern of its tail is similar to that of albinucha and fasciata. Meade- Waldo (1889) described it as having a very distinctive gait when TAXONOMY OF THE GENUS COLUMBA walking on the ground and a " soft, flopping flight " which was unlike that of any other pigeon known to him. He found that on the Island of Gomera, in the Canaries, it kept mainly to scrub-covered slopes at relatively low altitudes, being replaced in the high mountain forest by C. bollii. It lays one egg to a clutch. The Laurel Pigeon seems most likely to be either an offshoot from primitive palearctic Columba stock prior to its subsequent speciation or an early offshoot from palumbus. It is impossible to do more than guess at its affinities from an examination of skins and the little that has been recorded of its behaviour. A detailed comparison of its anatomical and behavioural characters with those of, at least, palumbus, fasciata and albinucha will probably be necessary before its true relationships can be ascer- tained. C. arquatrix Nominate arquatrix with a purple head, silver-grey nape and yellow bill and legs KEY TO SHADING Black or dark slate-grey Dark purple or dark purplish-chestnut Greyish-white or silver-grey Dark purple or purplish-chestnut with small white spots Dark purple with large pale grey spots Dark brown and purplish-brown Lanceolate display plumage with dark bases and shining tips to feathers FIG. 4. Diagrammatic sketches to show colours of the Arquatrix sub-group. Left to right : a. thomensis, a. sjostedti, a. arquatrix and (bottom) hodgsoni, polleni and albinucha. has always been considered, rightly in my opinion, conspecific with sjostedti from the Cameroon Mountains, which has an entirely dark grey head, red orange-tipped bill and dark purplish legs and feet. The Sao Thome representative, thomensis, is, however, currently given specific rank. It should, I think, be put as a race of arqua- trix. In plumage it is just a dull version of the latter with the spotting on wing- coverts and undersides and the lanceolate display plumage on the neck showing, apparently, the beginning of that obsolescence of markings that so often takes place in island forms. It otherwise shows characters of both mainland forms, having an 14 TAXONOMY OF THE GENUS COLUMBA entirely dark grey head like a. sjostedti but yellow legs and bill like a. arquatrix. It has a somewhat longer tail than either mainland form but this alone does not, in my opinion, warrant its specific separation from them. C. polleni* In spite of its almost uniform dark brown plumage polleni is also a representative of arquatrix. Although, presumably in the course of long isolation on the Comoro Islands, it has lost the rich colours of arquatrix it still retains lanceolate display plumage on the hind neck and the yellow legs and bill. Of three skins in the National Collection one shows faint pale spots on the underparts exactly similar to, though less pronounced than the spots of arquatrix. It has, however, reached a degree of differentiation from arquatrix that justifies, in my opinion, its specific rank, though it should be considered as forming a superspecies with arquatrix and hodgsoni. C. hodgsoni Verheyen (1955) considers this Indian species conspecific with arquatrix. That they are geographical representatives of the same stock cannot be doubted but hodgsoni differs from arquatrix in several minor points of coloration of plumage and soft parts and in its much greater sexual dimorphism. In view of this and of their wide geographic separation I think it is better to give them specific rank as members of the same superspecies. C. albinucha In coloration, colour-pattern and display plumage the White-naped Pigeon very closely resembles arquatrix. It differs in having red bill and feet, a patterned tail with conspicuous pale terminal bar and in lacking the white spots on the wing- coverts, although its underparts are spotted in the same manner as those of arquatrix. It is also a little smaller. The two are sympatric but their ecology and habitat preferences differ (Van Someren, 1949). Nothing seems to have been recorded of the displays, nesting or clutch-size of albinucha. Although there can be no doubt that they should be regarded as good species, the resemblances between albinucha and arquatrix are almost certainly due to close phylogenetic affinity and not to conver- gence. C. elphinstonii, C. pulchricollis and C. torringtoni Although differing in size and coloration these three forms all resemble each other in colour-pattern, form of display plumage and (so far as is recorded) in habits and ecology. It is difficult to decide whether it is best to treat them as races of a single * Through the kindness of Mr. C. W. Benson I have recently been able to examine the fine series of polleni collected on the B. O. U. Centenary Expedition to the Comoro Islands. These specimens show that the unsexed skins I had previously seen were all females and all somewhat "foxed". In fact males are a brownish purple on head, upper mantle and breast and the dark brown parts of both sexes are suffused with grey, especially on the rump. They thus resemble arquatrix more closely than I had thought and Mr. Benson informs me that the calls of the two are similar ; possibly identical. I. therefore, now consider that polleni should be treated as a race of arquatrix and not as a separate species. TAXONOMY OF THE GENUS COLUMBA 15 species or as members of a superspecies. Their ranges nowhere overlap and they have, presumably, been long isolated from contact with each other. As they have always been accorded specific rank it seems to me best to regard them as members of a superspecies. C. punicea The Purple Wood Pigeon most closely resembles argentina in colour-pattern although its purplish-chestnut body colour gives it a very different superficial appearance. This species is said sometimes to lay one and sometimes two eggs per clutch (Stuart Baker, 1913). If this is really the case no one other than Stuart Baker appears to have found nests with more than one egg this is an interesting intermediate condition between the species laying one egg and those laying two. C. argentina The Silver Pigeon shows a striking resemblance to the Pied Imperial Pigeons in colour and colour-pattern. Boetticher (1954) considered it more closely related to FIG. 5. Diagrammatic sketches to show colour-patterns of tails and outermost tail feathers of left, Ducula bicolor; right, Columba argentina. Ducula than to Columba. I have been unable to find any description of the anatomy of this pigeon which would show whether its stomach and intestines are of Ducula or Columba type. From its plumage characters, I think, however, that it is rightly placed in Columba. It shows the same type of sexual dimorphism female a slightly duller and browner grey in plumage colour as do other grey Columba species. It resembles punicea in the colour-pattern of its wings although the coloration of the two is different. Superficially the tail appears to be similarly marked to that of the Pied Imperial Pigeon but the pattern on the outer retrices (see Text-fig. 5) is quite different. I think, therefore, that argentina' s resemblance to the pied Ducula species is due to convergence and that it is best regarded as a rather aberrantly coloured member of the Elphinstonii species-group. It is another example of a Columba species which appears to have lost the display plumage on the neck, a point compatible with its presumed affinity to punicea which appears to show a regressional stage of this display plumage. C. palumboides In appearance, at least of the skin, the Andaman Wood Pigeon suggests a link between the Elphinstonii and Janthina species-groups. It seems likely to be nearer 16 TAXONOMY OF THE GENUS COLUMBA to the former on geographical grounds. Hume (1867) who collected some specimens and had a captive bird under observation was impressed by its " mode of holding itself and its broad substantial body " which he considered argued relationship to Ducula rather than to Columba. This may not be of importance as in so far as it is adapted to an arboreal life ; any large pigeon is bound to show some resemblance to Ducula as indeed does the common Wood Pigeon palumbus. A comparison of the alimentary tracts and a comprehensive study of the foods taken by this and other very arboreal Columba species with those of the Ducula species living in the same regions would be of great value. C. janthina The Black or Japanese Wood Pigeon and C. vitiensis, seem best considered as members of a single superspecies. C. j. nitens with its distinctive purple-brown head is somewhat intermediate, but in most respects is closer to nominate janthina than to any race of vitiensis. Stresemann (1939) suggests New Guinea as the place of origin of these forms and gives a map of their distribution and the probable directions of their former spread from the Papuan regions. C. vitiensis Although generally arboreal, so far as is known, C. v. halmaheira has been recorded nesting on the ground among thick cover (Warner, 1949) and Mr. I. C. J. Galbraith informs me that the natives on the Solomons told him that ground nesting was not infrequent in this species. In its bowing display (Newman, 1910) the bird moves rather slowly and does not erect its tail. C. norfolciensis This species appears to be a derivative of vitiensis. Males differ strikingly from the parent species in having the head and underparts white, or, in new plumage, creamy buff with a pink tinge. Females, however, have these parts whitish-grey or greyish-buff with some iridescence on the crown of the head and the pale throat as in vitiensis almost pure white and fairly clearly demarcated. Whether norfolciensis merits specific recognition is arguable but in view of the great difference in the appearance of the males, which makes it highly likely that if at some future period it were to invade the range of vitiensis or vice versa the two would not interbreed, and the nomenclatorial changes that would be involved, it seems preferable to accord it specific rank within the same superspecies. C. versicolor This extinct species appears to have been a very close relative of janthina. Kuroda (1925) considers it a race of janthina. Both forms occurred, however, on the Bonin Islands and I think therefore that Peters (1937) was right to give them specific rank in spite of their obvious affinities in form and coloration. TAXONOMY OF THE GENUS COLUMBA Q $ c'S O J3 I ll/ \ A o 2 ' I i 0> 4) 5 % g < 3 o I i 18 TAXONOMY OF THE GENUS COLUMBA C. jouyi Kuroda (1925) considers jouyi conspecific with janthina. Its distribution as given by him overlaps with nominate janthina on Okinawa. Jahn (1940) does not, however, record jouyi from Okinawa. Besides the large silver patch on its hind-neck and mantle which presumably functions as a recognition mark and probably also as an isolating mechanism -jouyi is a larger, longer-tailed bird and shows only slight traces of iridescence other than on the neck and nape. I therefore think it best to follow Peters in giving jouyi specific status. C. pallidiceps The Yellow-legged Pigeon appears to be a derivative from vitiensis stock. It now overlaps C. vitiensis halmaheira on some of the Solomon Islands (Mayr, 1934) although only pallidiceps occurs in New Britain, where it perhaps originated. Al- though much alike in general appearance the two differ strikingly in the coloration of the head, silver-grey in pallidiceps, dark purple with a white throat in vitiensis; and in the feet, which are red in vitiensis and yellow in pallidiceps. The coloration of pallidiceps is suggestive of that of norfolciensis (q.v.) whose close relationship to vitiensis is beyond doubt. It would be interesting to know if the difference in leg colour between the two serves in any way as an isolating mechanism. It would seem rather unlikely particularly as it is not correlated with any difference in bill colour, which is mainly red in both species. In other pairs of Columba species (e.g. arquatrix and albinucha) in which one has yellow and one red legs there is a correlated difference in bill colour. C. leucocephala The White-crowned Pigeon is very closely related to the Red-necked Pigeon C. squamosa. Both have the same highly-developed and beautifully (but differently) coloured display plumage on the back and sides of the neck, are of similar size and inhabit the same geographical areas. They may, however, be ecologically isolated as in Cuba (Barbour, 1943) ; leucocephala is a coastal and squamosa a highland form. The striking difference in coloration of the head, eyes and display plumage in these otherwise almost identical forms suggests that these differences function or have functioned as isolating mechanisms. It seems probable that this pigeon lays but one egg to a clutch (Wetmore and Swales, 1931) in spite of an earlier statement (Gundlach, 1874) that it lays two. C. squamosa The close relationship between the Red-necked and White-crowned pigeons has been discussed under leucocephala. The sexual dimorphism in squamosa is much less than in leucocephala but readily apparent when skins in a like state of plumage are compared. Probably this pigeon also lays only one egg (Wetmore, 1927) in spite of earlier statements (Gundlach, 1874) to the contrary. TAXONOMY OF THE GENUS COLUMBA 19 C. speciosa The Scaled Pigeon appears to be fairly close to sguamosa and leucocephala. In some respects its plumage is intermediate between that of these species and of picazuro but I do not agree with Verheyen that speciosa and picazuro should be generically separated from squamosa and leucocephala. In speciosa the dark edgings to the feathers are found throughout most of its plumage but are intensified in contrast and brilliance on the neck, thus giving a suggestion of the condition obtaining in leucocephala and squamosa. The Scaled Pigeon shows strong sexual dimorphism, the male having the mantle and wing-coverts rich purple (fading to chestnut) whereas those of the female are dull brown. C. picazuro The Picazuro Pigeon has the display plumage on the neck less highly developed than in leucocephala and squamosa but of an essentially similar kind. According to Hudson (1920) its habits, gait and behaviour are similar to those of palumbus but except for a vivid contrast of its gait on the ground with that of maculosa he gives no details. Hudson said that it lays two eggs, Venturi (see Hartert, 1909) that sometimes one and sometimes two eggs are laid. Wetmore (1926) found a nest with one egg. He also said that the display flight is like that of lima, a rather surprising fact in an essentially woodland species. C. corensis The Bare-eyed Pigeon can be considered as forming a superspecies together with picazuro. Its main differences, the much paler coloration and the very extensive amount of bare orbital skin, are probably adaptations to the hot, coastal areas where it lives. C. maculosa I think the Spotted Pigeon is most closely related to picazuro. It has evidently become more fully adapted to feeding on the ground in open country. Its resemblance in wing /tail proportions and gait to livia is almost certainly due to convergence in feeding habits. This is another species which has no display plumage on the neck, probably having lost it at a relatively recent date. If so, no close relationship to other species sharing this negative character is indicated. C. fasciata The Band-tailed Pigeon is reminiscent of C. arquatrix in its yellow bill and feet. The pattern of the bill is unusual in adult Columba but very much like that of Columba nestlings, only the dark sub-terminal band having " spread " to the tip of the bill and the lighter base turned bright yellow. The copulation ceremony of this species (Whitman, 1919) evidently differs in several details from that of European Columba species. 20 TAXONOMY OF THE GENUS COLUMBA C. araucana Although the Chilean Pigeon is very closely related to the southern form of the Band-tailed C . fasciata albilinea it has, I think, diverged sufficiently to be given specific rank. Besides the plumage differences (purplish-maroon on mantle and scapulars) it has a dark bill and red legs. It can be considered as forming a superspecies together with fasciata and caribaea. C. caribaea The Jamaican Band-tailed Pigeon is clearly derived either from fasciata, araucana or from some comparatively recent form ancestral to them both. In plumage it differs from fasciata only in being paler, in having lost the white neck-ring and having less well developed iridescent display plumage on the neck. It differs also in having, like araucana, a dark bill and red legs. C. cayennensis The Rufous Pigeon is intermediate in many respects between the members of the "fasciata " superspecies and those of the superspecies " flavirostris " . It resembles the former in the colour-pattern of its tail feathers and in having display plumage on the hind neck but in its other plumage characters it is closer to flavirostris . Little or nothing seems to be recorded about its breeding habits in the wild but two clutches of one egg each were laid by captive specimens (Newman, 1910). C. flavirostris The Red-billed Pigeon forms a superspecies together with oenops and inornata. All three completely lack display plumage on the neck thus presenting a superficial resemblance to the Oenoenas species. C. oenops This is the Peruvian representative of flavirostris. In appearance it differs only in the colour of its bill, which is yellow, or yellow tipped with black, instead of pink and whitish. Also the males have a larger area of ruddy purple on the plumage of the upper parts but females and juveniles are intermediate in this respect between males of their own species and flavirostris. C. inornata This island representative of flavirostris differs only in its larger size, black bill and grey, instead of red, orbital skin. The Sub-genus Oenoenas Salvadori (1893) characterized this sub-genus as having " general plumage, brown, more or less vinous, and with no light spots on the neck, but sometimes with concealed reddish spots, more or less obsolete, on the hind-neck ". Actually the " reddish spots " are appreciably lighter than the adjacent areas of the feathers TAXONOMY OF THE GENUS COLUMBA 21 that show them and these pigeons also have, as pointed out by Ridgway (1916), rather small bills and strongly rounded tails. Ridgway gave them, and others less deserving, full generic rank but subsequent authorities (Peters, 1937 ; Hellmayr & Conover, 1942 ; de Schaunsee, 1948) put Oenoenas into the synonymy of Columba. The four species concerned, plumbea, subvinacea, nigrirostris and goodsoni are all closely related and are more sharply differentiated than any others within Columba, mayeri excepted. I do not concur with Boetticher in considering them to be closely related to caribaea or that their dull concolorous plumage represents the retention of a primitive condition. The Oenoenas species differ from other Ameri- can pigeons (except for the quite differently coloured speciosa) in lacking grey on the rump, which is concolorous with the mantle and wings or nearly so. The more or less obsolescent pale or bronzy-red markings on the feathers of the hind-neck are more prominent in females than in males, which suggests that so far from repre- senting display plumage in process of acquisition, as Boetticher suggests, the reverse is the case. The obsolescent display plumage and general dull coloration of Oenoenas seems to be, in the evolutionary sense, a relatively recent development, not the retention of primitive characters. Oenoenas is most probably an early offshoot from the other American Columba that has diverged considerably from the parent stock. Its sombre, concolorous dress with complete lack of signal markings and obsolescent display plumage pre- sumably evolved as a consequence of increased predator pressure consequent on smaller size and /or increased reliance on vocal differences as isolating mechanisms. C. plumbea and C. subvinacea Only the fact that these pigeons appear to be sympatric in parts of their range, makes it possible on present knowledge to consider them as specifically distinct. Whilst the slight differences of colour and bill size make it possible to identify most specimens, it is difficult to believe that they could be sufficient to act as isolating mechanisms for the birds themselves. Presumably vocal and behavioural differences exist, but they do not appear to have been recorded. A few specimens in the National Collection are intermediate in character between plumbea and subvinacea but it is impossible to say whether these are hybrids or merely individuals whose slight aberrancy is towards the characters of the other species. C. nigrirostris and C. goodsoni These can certainly be regarded as forming a single superspecies. They may well be conspecific but since their taxonomic differences are at least as great (or one might better say no less) than those of plumbea and subvinacea, it seems better to grant them specific rank until more is known about them. The Sub-genus Turturoena This is comprised of three allopatric species delagorguei, iriditorques and malherbei which replace one another geographically and can be considered as forming a single superspecies. Salvadori (1893) separated Turturoena generically from Columba 22 TAXONOMY OF THE GENUS COLUMBA on its smaller size, hind neck with bright metallic colours and the sexes being often very dissimilar. Subsequent authorities have disagreed about its status, Peters (1937) placing it in the synonymy of Columba and Cave & Macdonald (1955) and Mackworth-Praed & Grant (1952) giving it full generic rank. I think, personally, that the differences do not merit more than sub-generic rank within Columba. Boetticher considers that Turturoena is more closely related to Aplopelia than to Columba. Certainly all three species have a similar distribution of iridescence to that of the Aplopelia species (but one found also in some species of Columba, and other pigeons) and a similar degree and type of sexual dimorphism. However, in my opinion, they show more affinities to the Columba species, not only in their general appearance but also in their colour-patterns which much resemble those found in palumbus and allied species. Mr. R. E. Moreau informs me (in litt.) that the cooing of delagorguei sharpei in Tanganyika reminded him very much of that of the Wood Pigeon, palumbus. The Sub-genus Nesoenas Salvador! (1893) placed the Mauritius Pigeon, C. mayeri, in the monotypic genus Nesoenas because of its more rounded wing, with the first primary about equal in length to the sixth, and its rufous tail. In coloration this pigeon does not closely resemble any other Columba but its colour-pattern shows some resemblance to the oenas species-group, from which it may possibly be an offshoot. Fortunately this species is not extinct as was at one time feared so it is possible that studies of it that will serve to indicate its affinities may yet be made. In the meantime I think it is better to put Nesoenas as a sub-genus within Columba than to afford it generic rank. SUMMARY This paper revises the genus Columba. The taxonomic and behavioural characters of the species are discussed in reference to their probable affinities. The absence or obsolescence of the display plumage on the neck is never a primitive character in Columba but one that has evolved independently in several species. Oenoenas, Turturoena and Nesoenas are given sub-generic rank. It does not seem feasible to further sub-divide the genus. There are indications that the American species may prove separable on anatomical and biochemical criteria although they cannot be separated on external characters. REFERENCES ALI, S. 1949. Indian Hill Birds. Oxford (p. 173). AUSTIN, O. L. 1948. The Birds of Korea. Bull. Mus. Comp. Zool. 101 (i) : 1-130. BAKER, E. C. STUART. 1913. Indian Pigeons and Doves. London. BARBOUR, E. 1943. Cuban Ornithology (Mem. Nuttall. Orn. Club. no. 9, pp. 70-72). BOETTICHER, H. VON. 1954. Die Taubengattung Columba L. Zool. Anzeiger, 153, 3-4 : 49-64. CAIN, A. J. 1954. Subdivisions of the genus Ptilinopus. Bull. Brit. Mus. (Nat. Hist.) Zool. 2, No. 8. TAXONOMY OF THE GENUS COLUMBA 23 CAVE, F. O. & MACDONALD, J. D. 1955. Birds of the Sudan. London. CUMLEY, R. W. & IRWIN, M. R. 1944- The correlation between antigenic composition and geographic range in the Old and New World of some species of Columba. Amer. Nat. LXVIII, No. 776 : 238-256. DEMENTIEV, G. P. et al. 1951. Birds of the Soviet Union, 11 : 6-32. GOODWIN, D. 1956. Remarks on the Rock and Speckled Pigeons. Bull. Brit. Orn. Club, 76 : 6-9. I956a. Observations on the voice and displays of certain pigeons. Avic. Mag. 62 : 17-33 and 62-70. GUNDLACH, J. 1874. Neue Beitrage zur Ornithologie Cubas. Journ. Orn. 1874 : 289. HARTERT, E. & VENTURI, S. 1909. Notes sur les oiseaux de la Republique Argentine. Novit. Zool. XVI, No. 2 : 260-261. HELLMAYR, C. E. & CONOVER, B. 1942. Catalogue of Birds of the Americas, pt. i, no. i, Zool. Ser. Field Mus. Nat. Hist. XIII, pt. i, No. i. HUDSON, W. H. 1920. Birds of La Plata. Vol. 2, pp. 154-155. JAHN, H. 1942. Zur Oekologie und Biologic der Vogel Japans. Journ. Orn. 90, 1-2 : 7-301. LA TOUCHE, J. D. D. 1934. -A handbook of the birds of Eastern China. Vol. 2, p. 209. London. KURODA, N. 1925. A contribution to the knowledge of the avifauna of the Riu Kiu Islands and the vicinity. Tokyo. MACKWORTH PRAED, C. W. & GRANT, C. H. B. 1952. The African handbook of Birds : Ser. i, vol. i, p. 468. MAYR, E. 1931. The birds of Malaita Island. Amer. Mus. Novit. 504, No. nth, 1931, p. 10. 1934- Notes on some birds from New Britain, Bismarck Archipelago. Ibid. 709, 4th April, 1934. PP- 4-5- MEADE- WALDO, E. G. 1889. Notes on some birds of the Canary Islands. Ibis, 6th Ser. No. i, pp. 6-9 and No. 4, pp. 509-510. MICULICZ-RADECKI, M. VON. 1949. Betrachtungen zur Stammesgeschichte der Wildtauben. Verhand. Deutsch. Zool., Aug. 1949, pp. 55-66. NEWMAN, T. H. 1910. Nesting of the White-throated Pigeon. Avic. Mag. 3rd Ser., vol. i, No. 5, pp. 158-164 and 193-195. 1911. The Snow Pigeon. Ibid. Vol. 2, No. 6, pp. 173-178. PETERS, J. L. 1937. Check-list of the Birds of the World. Vol. 3. Camb. Mass. U.S.A. POLLEN, P. L. & VAN DAM, D. C. 1868. Recherches sur la Faune de Madagascar et ser Dependences. Vol. 2, pp. 112-113. Leiden. SALVADORI, T. 1893. Catalogue of Birds in the British Museum. Vol. 21. SCHAFER, E. 1938. Ornithologische Ergebnisse zweier Forschungsreisen nach Tibet. Journ. f. Orn. Supplement, May 1938. SCHAUNSEE, R. M. DE. 1948. The Birds of the Republic of Columbia, part i. Bogota. SIBLEY, C. G. 1957. The evolutionary and taxonomic significance of sexual dimorphism and hybridisation in birds. Condor 59 : 3 : pp. 166-191. TAIBELL, A. M. 1954. Notizie sulla riproduzione in cattivat& del Colombo dal collare bianco (Columba albitorques}. Rivista It. Orn., 1954 (Seconda serie), pp. 195-203. VAN SOMEREN, V. G. L. 1949. The birds of Bwamba. Special supplement to the Uganda Journal, Vol. 13, pp. 24-25. VERHEYEN, R. 1955. Le Pigeon Bleu (Columba arquatrix Temm.) du Ruwenzori. Le Gerfaut, 1955. vol. 2, pp. 127-145. 1957. Analyse du potential morphologique et project de classification des columbiformes (Wetmore, 1934). Bui. Inst. Roy. Sci. Bel. Vol. 33, No. 3. WARNER, D. W. 1949. The White-throated Pigeon nesting on the ground in New Caledonia. Auk. 66 : 90-91. WETMORE, A. 1927. The birds of Porto Rico, pp. 390-391. WETMORE, A. & SWALES, B. H. 1931. The birds of Haiti and the Dominican Republic. U.S. Nat. Mus. Bull. 155 : 184. WHITMAN, C. O. 1919. The behaviour of pigeons. Posthumous works of C. O. Whitman. Vol. 3 : pp. 56-57. Carnegie Inst., Wash. ZOOL. 6, i. . r A REVISION OF THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) BASED ON MATERIAL IN THE COLLECTIONS OF THE BRITISH MUSEUM (NATURAL HISTORY) G. OWEN EVANS and K. H. HYATT BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 2 LONDON: 1960 A REVISION OF THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) BASED ON MATERIAL IN THE COLLECTIONS OF THE BRITISH MUSEUM (NATURAL HISTORY) BY G. OWEN EVANS and K. H. HYATT Department of Zoology, British Museum (Natural History) Pp. 25-101, 204 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6, No. 2 LONDON: 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 2 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued February, 1960 Price Twenty-four Shillings A REVISION OF THE PLATYSEIINAE (MESOSTIGMATA: ACEOSEJIDAE) BASED ON MATERIAL IN THE COLLECTIONS OF THE BRITISH MUSEUM (NATURAL HISTORY) By G. OWEN EVANS & K. H. HYATT CONTENTS INTRODUCTION EXTERNAL MORPHOLOGY CLASSIFICATION . Genus Platyseius Berlese Genus Plesiosejus gen. nov. Genus Sejus C. L. Koch Genus Zerconopsis Hull . SUMMARY .... REFERENCES Page 27 28 33 34 42 49 9i 100 IOO INTRODUCTION THE family Aceosejidae is divided into two subfamilies, Aceosejinae and the Platy- seiinae, on the basis of the chaetotaxy of the gnathosoma and the tarsi of legs II-IV (Evans, 1957). The classification of the Aceosejinae has been outlined by Evans (1958) but no revisionary work has yet been undertaken on the Platyseiinae which, at present, contains six genera, namely, Sejus C. L. Koch, 1843, Cheiroseius Berlese, 1916, Platyseius Berlese, 1916, Episeius Hull, 1918, Zerconopsis Hull, 1918, and Episeiella Willmann, 1938. Koch (1836) originally assigned a number of species to the genus Sejus but did not designate the type species of the genus, Sejus viduus C. L. Koch, until 1843. Berlese (1892) accepted Koch's wide concept of the genus and in 1913 proposed the family Seiidae for its reception together with eight other genera. Later in 1916, the same author revised his concept of the Seiidae and stated that Seius (= Sejus}, with Sejus togatus C. L. Koch as the type, was not confamilial with the other genera he had included in the family in 1913. Thus, the family Seiidae was reserved for those species having the general characteristics of Sejus togatus. Subsequent workers (Vitzthum and Tragardh) accepted Berlese's definition of Sejus and the family became established in the literature until 1936 when Oudemans noted the error in considering 5. togatus to be the type of Sejus. Consequently Sejus Berlese nee Koch became a synonym of Liroaspis Banks in spite of Tragardh's remark that " this incident offers yet another strong argument in favour of establishing nomina conservanda" (Tragardh, 1946). We are following Sellnick (see Willmann, 1953) ZOOL, 6. NO. 2. 3 28 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) who considers Epicrius corniger Berlese (1891) to be a synonym of Sejus viduus C. L. Koch. According to our generic concept Cheiroseius, Episeius and Episeiella are synonymous with Sejus. The present revision of the Platyseiinae is based on material in the Collections of the British Museum (Natural History). The major collections were made by the late A. H. G. Alston (Indonesia), P. F. Bellinger (Jamaica), Miss T. Clay (Sikkim), G. Owen Evans (Uganda), K. H. Hyatt (Nepal), P. Wygodzynsky (Argentina) and M. E. Bacchus, P. N. Lawrence and J. T. Salmon (British Isles). The type material, unless otherwise stated, is deposited in the British Museum (Natural History). EXTERNAL MORPHOLOGY Gnathosoma The chelicerae are chelate-dentate in all postembryonic developmental stages and show a marked uniformity in shape and dentition throughout the subfamily. The 1 2 3 FIG. 1-3. Chaetotaxy of the venter of the gnathosoma and palptrochanter, femur and genu in the larva (i), protonymph (2), and deutonymph (3) of Plesiosejus italicus (Berlese). i, rostral setae ; 2, external posterior rostral setae ; 3, internal posterior rostral setae ; 4, capitular setae ; n2, setae appearing in the deutonymph. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 29 8 FIGS. 4-9. Sejus necorniger (Ouderaans), dorsum of larva (4), protonymph (5), and deutonymph (6). Plesiosejus italicus (Berlese), dorsura of larva (7), protonymph (8), and deutonymph (9). 30 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) spermatophoral process on the movable digit in the male shows considerable inter- specific variation in form (e.g. Text-figs. 33 and 41) and appears to be a useful taxonomic character. In all stages the processes from the arthrodial membrane at the base of the movable digit are simple and setiform. The postembryonic development of the chaetotaxy of the venter of the gnathosoma and of the three basal movable segments of the pedipalp is typically parasitoid (Text-figs. 1-3). The rostral setae (i) in the immature stages and in the adults are long and usually whip-like as are the internal palptrochanter setae which appear in the deutonymph. (This characteristic structure of these two pairs of setae is a useful character for separating the platyseiines from the closely related aceosejines). The corniculi are simple and the specialized seta at the inner basal angle of the palptarsus is two pronged. With the exception of Sejus dayi sp. nov., the tectum is basically tridentate with the processes subequal in length and inconspicuously divided distally (Text-figs. 28, 32 and 76). In 5. dayi the median process is extremely long and Y- shaped with the lateral processes relatively shorter (Text-fig. 117). Idiosoma The idiosoma is usually elongate-oval in outline although in Platyseius it may be subcircular. Dor sum : the dorsum of the larva has two weakly sclerotized shields. In Sejus and Plesiosejus the anterior shield bears nine pairs of setae and the pygidial four pairs 1 (Text-figs. 4 arid 7). The striated cuticle between these sclerotized areas carries two pairs of J setae (Ji and J2), three pairs of Z setae (Zi-Z3) and one pair of S setae (? 83). Three or more platelets are also present on the cuticle. We have seen three main types of protonymph which chiefly differ in the chaetotaxy of the posterior half of the dorsum. All have an anterior shield with eleven or twelve pairs of setae and a pygidial shield. In the Sejus-type (Text-fig. 5) there are five pairs of setae in the J series ; Ji and J2 being situated on the cuticle between the anterior and pygidial shields. The protonymphs of both Platyseius and Plesioseius, on the other hand show a deficiency in the chaetotaxy of the J series. The former have only two pairs in the J series (J4 and J5) and the latter three pairs (J3-J5). There are no setae of the J series present on the cuticle between the anterior and pygidial shields in these genera (Text-fig. 8). All the deutonymphs we have examined have a laterally incised dorsal shield which almost entirely covers the dorsum of the mite. The chaetotaxy of the " anterior dorsal shield " (the region anterior to the incisions) is relatively constant but the number of setae on the "posterior shield" which is largely determined in the protonymph shows con- siderable variation. Four types may be readily recognized : i. Sejus-type : J, Z and S series each with five pairs of setae (Text-fig. 6). 1 The system of nomenclature for the dorsal chaetotaxy follows that proposed by Sellnick (1944) and added to by Hirschmann (1957). The latter author has used one system of nomenclature throughout the suborder and by so doing has attempted to homologize setae in widely differing groups. This may be possible in the larval stage but becomes increasingly difficult and purely subjective in subsequent developmental stages because of the increased number of setae, especially in those forms exhibiting hypertrichy. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 31 2. Zerconop sis-type : J and Z series each with five pairs, S series with four pairs. 3. Platyseius-type : J series with two pairs (J4 and J5), Z and S series each with five pairs. 4. Plesiosejus-type : J series with three pairs (J3-J5), Z and S series each with five pairs (Text-fig. 9). The dorsal shield in the adults is entire except in 5. clayi, Zerconopsis labradorensis sp. nov. and Zerconopsis muestari (Schweizer) which retain the deutonymphal incisions and Sejus ornatus sp. nov. which has wide incisions between 82 and 83 (Text-fig. 168). The chaetotaxy of the " posterior dorsal shield " is essentially the same as in the deutonymph so that the above four groups may also be recognized in the adult. The surface of the shield is invariably reticulated and the majority FIGS. lo-n. 10 Dorsal shield of Plesiosejus italicus (Berlese). transverse section. Fig. 10, lateral view. Fig. n, of the species have a median longitudinal ridge (Text-figs. 10 and n). The position of the highest point along this ridge varies throughout the group. It is usually located in the region of setae I4-J2 but in some species it may appear as a distinct tubercle between J4 and J5. The extreme attenuation of the posterior margin of the anterior shield in the protonymph of Plesiosejus italicus (Berl.) may be due to the development of the median ridge (cf. Text-figs. 8 and 36). Venter : The tritosternum is well developed in all stages and in the adult comprises a long slender base and a pair of strongly pilose laciniae. Pre-endopodal shields may be present in some species. The development of the chaetotaxy of the intercoxal region is normal ; the larva having three pairs (homologous with sternal setae I-III in the adult), the proto- nymph four pairs (homologous with sternal setae I-III and the genital setae in the adult) and the deutonymph five pairs (the extra setae appearing at this stage being homologous with the metasternals in the adult). The female has a strongly sclero- tized sternal shield bearing three pairs of setae and the metasternals are normally situated on discrete platelets. The genital shield is wedge shaped and except in S. clayi and certain species of Zerconopsis has a pair of setae. A sterniti-genital shield is always present in the 32 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) male although it may show various degrees of fusion with the ventral shield. The genital orifice in this sex is located near the anterior margin of the sterniti-genital. The larva has an anal shield usually with a seta on each anal valve in addition to the normal three setae present on the shield. There is a tendency for increased scJtrotization of the anal region in subsequent developmental stages so that the protonymph and the deutonymph may have a ventri-anal shield (an anal shield with usually one pair of preanals). All the females of the platyseiines we have examined have a ventri-anal shield with from one to four pairs of pre-anals. Mem- bers of the genus Platyseius in addition have two or more pairs of setae situated in the posterior half of the shield (Text-fig. 23). The ventri-anal shield in the male shows varying degrees of fusion with the sterniti-genital shield and with the dorsal shield (Text-figs. 19 and 31). FIGS. 12-14. Tarsi of leg I showing development of ambulacra. Fig. 12, Sejus viduus C. L. Koch. Fig. 13, Sejus laelaptoides (Berlese). Fig. 14, Sejus unguiculatus (Berlese). The larva lacks stigmata and peritremes but the protonymph has a pair of stigmata located ventro-laterally between coxae III and IV and short peritremes. In the deutonymph, the peritremes are well developed and extend beyond the anterior margin of coxae I. The peritrematal shields at this stage, however, are weak and inconspicuous. The adults have large peritrematal shields ; these are invariably fused with the expodals and extend posterior to coxae IV (Text-fig. 16). Many species have a strong poststigmatic prolongation of the peritreme. Legs : All the legs are usually long and slender. The structures of taxonomic importance are the form of the ambulacrum and the chaetotaxy of tarsi II-IV. The ambulacrum of leg I shows varying degrees of development. In a few species it is lacking (Text-fig. 12) but generally it comprises a lobate pulvillus and two claws. The claws may be minute or relatively large and conspicuous (Text-fig. 13). A short pretarsus is usually present ; a notable exception being Sejus unguiculatus (Berl.) in which the large claws on tarsus I are sessile (Text-fig. 14). The ambulacra THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 33 of legs II-IV are well developed and each consists of a pair of claws and a lobate pulvillus. In the majority of the platyseiines, the median and lateral lobes of the pulvilli are slender and markedly acute a probable adaptation for movement over a moist substratum. Tarsi II and III and usually IV have a pair of characteristic lanceolate setae (Text-fig. 18). Leg II in the male is never crassate or armed with spurs. CLASSIFICATION The characters previously used in separating the genera of the Platyseiinae, e.g. the outline of the idiosoma, the presence or absence of an ambulacrum on tarsus I and the form of the peritreme, are undoubtedly of some value in the practical classification of the subfamily but have no phylogenetic basis. We have concluded from our investigations of the external morphology of the group that the chaetotaxy and form of the dorsal shield in the adult and immature stages afford the most satisfactory characters for a natural classification. It is interesting to note that the variation in the form and in the chaetotaxy of the dorsal shield already noted for the Aceosejinae (Evans, 1958) is also evident in the Platyseiinae. Thus, there is a tendency towards the reduction in the number of the J and S setae on the " posterior dorsal shield " and towards the retention of the deutonymphal incisions. Our generic concept has as its basis the chaetotaxy of the " posterior dorsal shield " which has been discussed above. Subfamily PLATYSEIINAE Platyseiinae Evans, G. O., (1957), / Linn. Soc. Land. (Zool.), 43 : 244. Aceosejids with the rostral and internal palptrochanter setae long and usually whip-like. Tarsi II and III with a pair of long lanceolate setae about the middle of the segment (Text-fig. 18). Dorsal shield in the adults entire, rarely with lateral incisions. " Posterior dorsal shield" normally with fifteen pairs of setae; when fewer setae are present the J series or S series comprises less than five pairs. Females with sternal shield bearing three pairs of setae, a wedge-shaped genital shield with or without a pair of setae and a ventri-anal shield with one or more pairs of preanals. Male with sterniti-genital and ventri-anal shields, the latter often being partially fused with the dorsal shield. Both sexes with or without poststigmatic prolongation of the peritreme. Chelicerae chelate-dentate in both sexes ; movable digit with spermatophoral process in the male. Tectum basically three pronged. All legs usually slender ; leg I with or without ambulacrum, legs II-IV with pulvilli (often acutely lobate) and two claws. Leg II unarmed in the male. KEY TO GENERA ADULTS i. " Posterior dorsal shield " with 5 pairs of setae in the J series (Text-fig. 53) . " Posterior dorsal shield " with only 2 or 3 pairs of setae in the J series (Text-figs. 15 and 36) .,,,,,,, i .... 34 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 2. " Posterior dorsal shield " with 5 pairs of setae in the S series (Text-fig. 53) ; dorsum never with paddle-like setae ; peritreme often with a strong poststigmatic process (Text-fig. 54) ; median lobes of the pulvilli of ambulacra II-IV usually acuminate (Text-fig. 58) Sejus C. L. Koch (p. 49) " Posterior dorsal shield " with 4 pairs of setae in the S series, ? 82 situated on the lateral integument (Text-fig. 179) ; dorsal shield with three or more pairs of paddle-like setae, remainder of dorsal setae simple (Text-fig. 184) ; peritreme without poststigmatic prolongation ; median lobes of the pulvilli of legs II-IV rounded apically (Text-fig. 202) . . . . Zerconopsis Hull (p. 91) 3. Both sexes with three pairs of setae in the J series (Text-fig. 36) ; idiosoma elongate- oval in outline ; ventri-anal shield in the female with 3 or 4 pairs of setae in addition to the paranal and postanal setae . . . Plesiosejus gen. nov. (p. 42) Both sexes with only 2 pairs of setae in the J series (J4 and Js) ; idiosoma usually subcircular in outline (Text-fig. 15) ; ventri-anal shield in the female with 5 or 6 pairs of setae in addition to paranal and postanal setae . Platyseius Berlese (p. 34) Genus PLATYSEIUS Berlese Platyseius Berlese, A., (1916). Redia, 12 : 42. Platyseiine mites with the idiosoma usually subcircular in outline. Dorsal shield entire : " posterior dorsal shield " with J series comprising two pairs of setae (J4 and J5). Dorsal setae, with the exception of J5 in some species, long and simple. Sternal shield in the female with three pairs of setae ; metasternals situated on small plates. Genital shield wedge-shaped and with a pair of setae. Ventri-anal shield large and bearing 13 or 15 setae. Poststigmatic process of peritreme and peritrematal shield well developed and extending posterior to coxa IV. Male with sterniti-genital and ventri-anal shield, and ventri-anal shield showing varying degrees of fusion with the dorsal shield. Chaetotaxy of gnathosoma normal. Chelicerae chelate-dentate ; movable digit with spermatophoral process in the male. Tectum basically trispinate. All legs with ambulacra ; pulvilli of ambulacra II-IV acuminate. Type species Lasioseius (Platyseius) capillatus Berlese, 1916 (=Hypoaspis subglabra Oudemans, 1902) KEY TO SPECIES FEMALES Setae 15 shorter than J4 (Text-fig. 15) ; ventri-anal shield with 13 setae ; pre- endopodal shields present. Europe . . . ^ P. subglaber (Oudms.) (p. 35) Setae i5 at least twice the length of J4 (Text-fig. 22) . . . . . 2 Ventri-anal shield with 15 setae (Text-fig. 27) ; pre-endopodal shields absent. Jamaica ........ P. jamaicensis sp. nov. (p. 39) Ventri-anal shield with 13 pairs of setae (Text-fig. 23) ; pre-endopodal shields present ............. 3 Setae on ventri-anal shield simple (Text-fig. 23) ; setae J5 thorn-like, smooth. Indonesia ......... P. mollicomus Berl. (p. 37) Two pairs of setae on the posterior half of the ventri-anal shield stout, spinose (Text-fig. 35) ; setae J$ short, brush-like. East Africa P- spinosus sp. nov. (p. 41) THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) Platyseius subglaber (Oudemans), 1903 35 Hypoaspis subglabra Oudemans, A. C., (1903). Ent. Ber., 1 : 87 ; Abh. nat. Ver. Bremen, 18 : 88. Lasioseius (Platyseius) capillatus Berlese, A., (1916). Redia, 12 : 43, syn. nov. Lasioseius (Platyseius) subglaber, Halbert, J. N., (1923). /. Linn. Soc. Lond. (Zool.), 35 : 373. Lasiosei^is (?) listrophorus Schweizer, J., (1924). Arch fur Hydrobiol., 15: 131. FEMALE. Dorsal shield (560 ju- x 361-477 ju,) broadly oval ; ornamentation restricted to a number of small punctate areas. " Anterior dorsal shield " with FIGS. 15-18. Platyseius subglaber (Oudemans), female. Fig. 15, dorsum. Fig. 16, venter. Fig. 17, chelicera. Fig. 18, tarsus II. nineteen pairs of simple setae ; verticals very fine and short (Text-fig. 15). " Posterior dorsal shield " with twelve pairs of setae ; Ji-j3 being absent. Setae J5 short, palmate. The distribution and relative lengths of the setae on the dorsal shield and lateral interscutal membrane are shown in the figure. 36 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Tritosternum large with a pair of strongly-pilose laciniae, and flanked at its base by two pairs of pre-endopodal shields. Sternal shield smooth and with three pairs of simple setae (Text-fig. 16). Genital shield wedge shaped and with a pair of setae. Ventri-anal shield wider than long (202-215 /i long X 280-360 IJL wide), reticulated in its anterior half and bearing thirteen setae. Six small sclerotized plates are situated between the genital and ventri-anal shields. Stigma situated FIGS. 19-21. Platyseius subglaber (Oudemans), male. Fig. 19, venter. Fig. 21, tarsus II. Fig. 20, chelicera. between coxae III and IV ; peritreme extending posterior to the stigma ; peri- trematal shield strongly developed. Metapodal shields apparently fused with the peritrematal shields. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip-like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. Dentition of chelicerae as in Text-fig. 17. THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) 37 Leg I (670-772 fi long) with the tarsus (160-177 /*) considerably longer than the tibia (126-140 //). Setae on tibia and tarsus fine, those on remaining segments stouter. Tarsi II-IV with a pair of long lanceolate setae. Ambulacrum of tarsus I short and claws small. Pulvilli of ambulacra II-IV produced into three long acuminate lobes (Text-fig. 18). MALE. Dorsal shield (425-467 p X 340-371 /*) is similar to that of the female. " Anterior dorsal shield " with nineteen pairs of simple setae and " posterior dorsal shield " with twelve pairs of setae. The tritosternum is flanked at its base by two or three pairs of small pre-endopodal shields. The remainder of the chaetotaxy and sclerotization of the venter is shown in Text-fig. 19. The gnathosoma and tectum are similar to those of the female. The sperma- tophoral process is about three times as long as the movable digit of the chelicera. The dentition of the chelicera, and the spermatophoral process are shown in Text- fig. 20. Leg I (602-682 /* long) has the tarsus (147-177 /*) considerably longer than the tibia (114-126 fi). The chaetotaxy of the legs is as in the female. The ambulacra of tarsi II-IV are provided with lanceolate pulvilli (Text-fig. 21). DISTRIBUTION. This species is widely distributed in western Europe. In Great Britain we have examined specimens from Cumber Park, Nottinghamshire (coll. G. O. E., April, 1955), Dungeness, Kent (coll. D. A. Chant, July, 1955), Box Hill, Surrey (coll. E. Duffey, 30. vi. 1951), and St. Agnes, Isles of Stilly (coll. K. H. H., September, 1957). Platyseius mollicomus Berlese, 1916 Lasioseius (Platyseius} mollicomus Berlese, A., (1916). Redia, 12 : 43. Platyseius mollicomus, Vitzthum, H. G., (1931). Arch, fur Hydrobiol., Suppl., 9 : 70, figs. FEMALE. Dorsal shield (570-580 ju, X 467-480 /*) broadly oval in outline ; ornamentation restricted to a number of faint reticulated areas. " Anterior dorsal shield " with twenty-one pairs of simple setae (Text-fig. 22). " Posterior dorsal shield " with twelve pairs of simple setae ; Ji-Js being absent. Setae J5 are short and simple. The distribution and relative lengths of the setae on the dorsal shield are shown in the Text-fig. Tritosternum large with a pair of strongly pilose laciniae, and flanked at its base by two pairs of pre-endopodal shields. Sternal shield with three pairs of simple setae and faintly reticulated (Text-fig. 23). Genital shield slightly convex pos- teriorly and with a pair of setae. Ventri-anal shield wider than long (177 // long X 318 IL wide) with a transverse line anteriorly, and bearing thirteen setae. Four small sclerotized plates are situated between the genital and ventri-anal shields. Stigma situated between coxae III and IV. Peritreme broad, occupying most of the peritrematal shield, and with two deep indentations (Text-fig. 24), and extending posterior to the stigma. Metapodal shields fused with the peritrematal shields. Venter of gnathosoma with rostral and internal palptrochanter setae long and 38 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) whip-like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The chelicerae are typical. Leg I (700-760 fi long) with the tarsus (202 fi) longer than the tibia (126 //). Setae on tarsus fine, those on remaining segments stouter. Tarsi II-IV with a pair of long lanceolate setae. Ambulacrum of tarsus I fairly short and claws small. Pulvilli of ambulacra II-IV produced into three acuminate lobes (Text-fig. 25). FIGS. 22-25. Platyseius mollicomus Berlese, female. Fig. 22, dorsum. Fig. 23, venter. Fig. 24, peritrematal shield. Fig. 25, tarsus II. MALE. This sex (415 /i x 340 ju,) is described by Vitzthum (1931) who also gives a figure of the chelicera. DISTRIBUTION. Berlese (1916) based his description on material collected in Java by Cl. Jacobson. Vitzthum (1931) examined specimens (male, female and deutonymph) from Tjibodas, Java (Indonesia). The Museum Collections contain a female and a deutonymph collected by A. H. G. Alston in decaying leaves and THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) 39 fruit, Bogor, Indonesia (1956.2.27.6) ; and a single female collected by G. Owen Evans in grassland nine miles north-east of Bundibugyo, Ruwenzori, Uganda, 24.viii.i952 (1959.1.20.1). Platyseius jamaicensis sp. nov. FEMALE. Dorsal shield (498 ju, x 392-402 /*) broadly oval in outline ; ornamenta- tion restricted to a number of small punctate areas. "Anterior dorsal shield" with twenty pairs of simple setae of which the verticals are the shortest (Text-fig. 30 FIGS. 26-30 Platyseius jamaicensis sp. nov., female. Fig. 26, dorsum. Fig. 27, venter. Fig. 28, tectum. Fig. 29, chelicera. Fig. 30, tarsus II. 26). " Posterior dorsal shield " with thirteen pairs of setae ; Ji-j3 being absent. Setae J5 short, palmate and inconspicuous. J4 and Z5 are stout. The distribution and relative lengths of the setae on the dorsal shield are shown in the Text-fig. Tritosternum large with a pair of strongly-pilose laciniae. Sternal shield with a pair of longitudinal elongate punctate areas and three pairs of simple setae (Text- fig. 27). Genital shield convex posteriorly and finely reticulated, with a pair of setae. Ventri-anal shield wider than long (152 /i long X 253-273 p wide), reticulated 40 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) anteriorly and bearing fifteen setae. Four small sclerotized plates are situated between the genital and ventri-anal shields. Stigma situated between coxae III and IV, peritreme extending posterior to the stigma ; peritrematal shield strongly developed. Metapodal shields apparently fused with the peritrematal shields. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip-like. All setae on pedipalp simple. Tectum (Text-fig. 28) three-pronged FIGS. 31-33. Platyseius jamaicensis sp. nov., male. Fig. 31, venter. Fig. 32, tectum. Fig. 33, chelicera. and typical of the genus. Dentition of chelicera as in Text-fig. 29. Leg I (about 770 /i long) with the tarsus (177 p] longer than the tibia (157 /*). Setae on tibia and tarsus very fine : those on remaining segments stouter. Tarsi II-IV with a pair of long lanceolate setae. Ambulacrum of tarsus I short and the claws small. Pulvilli of ambulacra II-IV produced into three acuminate lobes (Text-fig. 30). THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 41 MALE. Dorsal shield (498 /i x 310 ft) similar to that of the female. The tri- tosternum is of the same form as in the female. The chaetotaxy and sclerotization of the venter are shown in Text-fig. 31. The gnathosoma is similar to that of the female, and the tectum is shown in Text-fig. 32. The spermatophoral process is about twice as long as the movable digit of the chelicera (Text-fig. 33). Leg I (c. 680 ft) with the tarsus and tibia each 152 ft. The chaetotaxy and ambulacra of the legs are as in the female. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.18) and one paratype from Fairy Glade, St. Andrew, in Lycopodium, 4,300 ft., 6.xii.i956; and the allotype male (1959.1.20.19) from St. Thomas, John Cove Mountains, in moss, 2,300 ft., 25. v. 1956 collected by P. F. Bellinger. The paratype female has been returned to the collector. Platyseius spinosus sp. nov. FEMALE. Dorsal shield (498 /* x 402 (JL) broadly oval in outline ; ornamentation restricted to a number of very faint reticulated areas. " Anterior dorsal shield " FIGS. 34-35. Platyseius spinosus sp. nov., female. Fig. 34, dorsal shield. Fig. 35, venter. with twenty-one pairs of simple setae (Text-fig. 34). Posterior dorsal shield with thirteen pairs of simple setae ; Ji-Js being absent. Setae J5 are short, palmate and inconspicuous. Z5 are stout. The distribution and relative lengths of the setae on the dorsal shield are shown in the Text-fig. \ ZOOL. 6, NO. 2. 42 THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) Tritosternum large with a pair of strongly-pilose laciniae, and flanked at its base by one or two pairs of pre-endopodal shields. Sternal shield with three pairs of simple setae and with a characteristically-shaped central mark (Text-fig. 35). Genital shield slightly convex posteriorly and with a pair of setae. Ventri-anal shield wider than long (177 p long x 298 p wide) with a transverse line anteriorly and bearing thirteen setae of which two pairs are stout. A narrow sclerotized band and two narrow plates are situated between the genital and ventri-anal shields. Stigma situated between coxae III and IV, peritreme broad, occupying most of the peritrematal shield, with a deep indentation, and extending posterior to the stigma. Metapodal shields fused with the peritrematal shields. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The chelicerae are typical. Leg I (707 /* long) with the tarsus (202 fi) longer than the tibia (126 fi). Setae on tarsus fine ; those on remaining segments stouter. Tarsi II-IV with a pair of long lanceolate setae. Ambulacrum of tarsus I short and claws small. Pulvilli of ambulacra II-IV produced into three acuminate lobes. MALE. Unknown. LOCALITY. Ruwenzori, Uganda. A single female (1959.1.20.20), collected by G. Owen Evans, in grassland at 2,850 ft., 24.viii.i952, nine miles north-east of Bundibugyo. Genus PLESIOSEJUS gen. nov. Platyseiine mites with the " posterior dorsal shield " having only three pairs of setae in the J series (Ji and J2 absent). Z and J series each with five pairs. Lateral margins of the dorsal shield entire in the adult stages. Vertical setae present or absent. Sternal shield in the female with three pairs of simple setae ; metasternals free. Genital shield wedge-shaped and bearing a pair of setae. Ventri-anal shield large with three or four pairs of setae in addition to the paranals and the postanal seta. Male with sterniti-genital and ventri-anal shields, the latter showing varying degrees of fusion with the podal-peritrematal and dorsal shields. Type species Ameroseius it aliens Berlese, 1905 KEY TO SPECIES FEMALES Without poststigmatic prolongation of peritreme (Text-fig. 50) ; majority of dorsal setae long, stout (Text-fig. 49) ; vertical setae well developed, directed anteriorly ; ventri-anal shield with 9 setae. Jamaica . . P. horridus sp. nov. (p. 48) With strong poststigmatic prolongation of peritreme extending posterior to coxa IV ; dorsal setae considerably shorter (Text-fig. 36) ; vertical setae absent ; ventri-anal shield with 1 1 setae ......... 2 Dorsal shield with distinct " tubercle " posterior to J4 (Text-fig. 36). Europe P. italicus (Berlese), (p. 43) Dorsal shield without such " tubercle " (Text-fig. 42). P. major (Halbert), (p. 47) THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 43 Plesiosejus italicus (Berlese), comb. nov. Ameroseius italicus Berlese, A., (1905). Redia, 2 : 234 ; (1906). Ibid., 6 : tav. 19, Text-figs. 35. 35a. Paraseius tenuipes Halbert, J. N. (1915). Proc. R. Irish Acad., 39ii : 78, syn. nov. Lasioseius italicus Berlese, A., (1916). Redia, 12 : 34. i Lasioseius (Episeius) michaeli Halbert, J. N., (1923). /. Linn. Soc. Lond. (Zool.), 35 : 373. Episeius tenuipes, Schweizer, J., (1949). Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 60. E. tenuipes, Franz, H., (1943). Denkschr. Akad. Wiss. Wien., 107 : 87. FEMALE. Dorsal shield (550-651 p long x 310-402 ju, wide) oval; ornamentation comprising a reticulated pattern covering the whole shield (Text-fig. 36). " Anterior dorsal shield " with twenty pairs of simple setae ; verticals absent. " Posterior dorsal shield" with thirteen pairs of setae; Ji-j2 being absent; J5 are finely pilose. A strong tubercle arising posterior to setae J3 denotes the highest point on the median ridge of the dorsal shield. The distribution and relative lengths of the setae and the structure of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae, and flanked at its base by one or two pairs of small pre-endopodal shields. Sternal shield with three pairs of simple setae and two oval areas antero-medially (Text-fig. 37). Genital shield convex posteriorly, and bearing one pair of setae. Ventri-anal shield considerably wider than long (177-202 /* long X 265-341 ft, wide), reticulated and bearing eleven simple setae. Between the genital and ventri-anal shields lie six narrow sclerotized plates. Stigma between coxae III and IV and the peritreme extending posterior to coxa IV ; peritrematal shield strongly sclerotized. Metapodal shields free and transversely situated. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 38. Leg I (639-726 ft long) with the tarsus (167-195 ju) longer than the tibia (113- 129 fi). Setae on tarsus and tibia very fine ; those on remaining segments stouter ; ambulacra short and claws small. Tarsi II-IV with a pair of lanceolate setae ; pulvilli produced into three acuminate lobes (Text-fig. 39). MALE. Dorsal shield (467-507 ju, long x 279-310 fi wide) entirely reticulated as in the female. "Anterior dorsal shield" with nineteen pairs of simple setae, verticals absent. " Posterior dorsal shield " with thirteen pairs of setae ; Ji-Js absent ; J5 short and pilose. Postero-dorsal tubercle not so pronounced as in the female. Tritosternum as in the female. The chaetotaxy and sclerotization of the venter are shown in Text-fig. 40. The gnathosoma and tectum are similar to those of the female. Spermatophoral process four times as long as the movable digit of the chelicera (Text-fig. 41). Leg I (534 ju, long) with the tarsus (147-152 ft) longer than the tibia (96-106 /*). The chaetotaxy and ambulacra of the legs are similar to the female. 44 THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) DISTRIBUTION. Berlese's (1905 and 1916) descriptions of the male and female of italicus are based on specimens collected under dead rotting leaves on a marsh at Udine in northern Italy. Halbert (1915) described P. tenuipes from a single female " found amongst moss on a stone in a mountain stream at Glencree, County Dublin, May ". Schweizer (1949) records tenuipes from the Swiss National Park. We have FIGS. 36-39. Plesiosejus italicus (Berlese), female. Fig. 36, dorsum. Fig. 37, venter. Fig. 38. chelicera. Fig. 39, tarsus II THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSE J ID AE) 45 examined British specimens from the following localities : Thirlmere, Cumberland, 25. ix. 1956, coll. P. N. Lawrence ; Grasmere, Westmorland, 29.^.1951, coll. J. T. Salmon; Blelham Tarn, Lancashire, 22.iii.i955, coll. R. J. Elliott; summit of Snowdon, Caernarvonshire, August, 1951, coll. E. Duffey ; Minworth, Birmingham, io.ix.T-952, coll. T. G. Tomlinson ; Swithland Wood and Anstey, Leicestershire, 19. iv. 1958, coll. P. N. L.; Sea Houses, Northumberland, i.x.i95i, coll. J. T. FIGS. 40-41. Plesiosejus italicus (Berlese), male. Fig. 40, venter. Fig. 41, chelicera. Salmon; Oxted, Surrey, 20. i. 1957, coll. P. N. L.; Leith Hill, Surrey, 6.vi.i95i, coll. J. T. Salmon; Keston, Kent, 2.viii.i957, coll. P. N. L. ; Dungeness, Kent, S.vii.igss, coll. D. A. Chant; Avonmouth, Gloucestershire, ii.v.1957, coll. P. N. L.; Combe Martin, Devonshire, March and June, 1957, coll. P. N. L. and M. E. Bacchus ; Forth Gwarra, Cornwall (A. D. Michael Coll.) ; The Burren, Co. Clare, 6.vi.i95i, coll. E. A. J. Duffy. 4 6 THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) FIGS. 42-47. Plesiosejus major (Halbert). Fig. 42, dorsum of female. Fig. 43, venter of female. Fig. 44, chelicera of female. Fig. 45, tarsus II of female. Fig. 46, venter of male. Fig. 47, chelicera of male. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 47 Plesiosejus major (Halbert), comb. nov. Lasioseius (Episeius) major Halbert, J. N., (1923). /. Linn. Soc. Zool. Land. 35 : 373. Lasioseius (L.) tenuipes, Schweizer, J., (1922). Verh. Naturf. Ges. Basel, 33 : 43. Episeius major, Schweizer, J., (1949). Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 61. FEMALE. Dorsal shield (581-740 ju, long x 371-488 fi wide) broadly oval in outline : ornamentation consisting of a dense reticulated pattern covering the whole of the shield (Text-fig. 42). " Anterior dorsal shield " with twenty pairs of simple setae. " Posterior dorsal shield " with thirteen pairs of setae ; Ji-j2 being absent ; setae J5 are short and spiculate. The distribution and relative lengths of the dorsal setae are shown in the Text-fig. Tritosternum well developed with pilose laciniae, and flanked at its base by two or three pairs of small pre-endopodal shields. Sternal shield with three pairs of simple setae (Text-fig. 43). Genital shield with one pair of setae and slightly convex posteriorly. Ventri-anal shield wider than long (203-253 ju, long X 288-379 /* wide), strongly reticulated and bearing eleven simple setae. Six narrow sclerotized plates are situated transversely on the membrane between the genital and ventri-anal shields. Stigma situated between coxae III and IV with the peritreme extending posterior to coxa IV. Peritrematal shield strongly sclerotized. Metapodal shields free and obliquely situated. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 44. Leg I (681-797 /* l n g) with the tarsus (147-177 p] longer than the tibia (106- 152 fi) ; pulvillus short, claws small. Setae on tarsus and tibia very fine ; those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae ; pulvilli produced into three acuminate lobes (Text-fig. 45). MALE. Dorsal shield (550 fi long x 340 /* wide) of the same form and chaetotaxy as the female. The structure and chaetotaxy of the venter are shown in Text-fig. 46. The gnathosoma and tectum are similar to those of the female. The sperma- tophoral process is twice as long as the movable digit of the chelicera (Text-fig. 47). Leg I (670 ju, long) with the tarsus (165 /*) longer than the tibia (126 fi}. The legs are similar to those of the female. DISTRIBUTION. Halbert's (1923) description of males and females of this species is based on specimens found in moss in streams at Kilmashogue and Glencullen in Ireland, and at Mill Bay, Land's End, England. Schweizer (1922) described what he considered to be males of Lasioseius tenuipes (Halbert) from Basel, Switzerland, but his figures show clearly that his specimens are Plesiosejus major (Halbert). In 1949 the same author recorded major from the Swiss National Park. We have examined specimens collected by the River Tolka, Glasnevin, Co. Dublin (Halbert Collection), and from the following localities: Oxted, Surrey, 20. i. 1957, coll. P. N. Lawrence ; Cumber Park, Nottinghamshire, April, 1955, coll. G. O. E. ; Markfield and Anstey, Leicestershire, 19. iv. 1958, coll. P. N. L.; Blea Tarn, West- 4 8 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) morland, 27. ix. 1956, coll. P. N. L.; and Chorley, Lancashire, 21. xi. 1940, coll. Water Pollution Research Laboratory. All these British specimens were collected from wet habitats (mosses, etc.). Plesiosejus horridus sp. nov. FEMALE. Dorsal shield (350 /* long x 218 fi wide) narrower in outline than in the other species of the genus, and is steeply elevated posteriorly (Text-fig. 48). FIGS. 48-52. Plesiosejus horridus sp. nov., female. Fig. 48, lateral view of dorsum. Fig. 49, dorsal shield. Fig. 50, venter. Fig. 51, chelicera. Fig. 52, tarsus II. There is no noticeable ornamentation on the dorsal shield. " Anterior dorsal shield" with twenty pairs of setae. Ji-Js absent. Setae J5 are partly concealed by the postero-dorsal elevation of the dorsal shield which bears J4 at its apex (Text-fig. 49). The distribution and relative lengths of the dorsal setae are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with three pairs THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 49 of short simple setae (Text-fig. 50). Genital shield flask shaped with a pair of setae. Ventri-anal shield slightly wider than long (93 fi long X 113 [i wide), with nine setae of which two lateral pairs are long. There appears to be a narrow sclerotized band between the genital and ventri-anal shields. Stigma between coxae III and IV, with the narrow, crenate peritreme not extending posterior to the stigma. Peri- trematal shield extending to the posterior margin of coxa IV. Metapodal plates oval, free. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 51. Leg I (370 (i long) with the tarsus (96 fi) one-and-a-half times as long as the tibia (63^). Setae on tarsus very fine ; those on remaining segments stouter. Tarsi II- IV with a pair of lanceolate setae. Tarsus I without claws. Pulvilli of tarsi II-IV produced into three acuminate lobes (Text-fig. 52). Leg IV 370 ju, long with the tarsus 130 fi. MALE. Unknown. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.41) and two paratype females (1959.1.20.42) in damp leaf litter, St. Ann, Mt. Diablo, 2,200 ft., 3.vi.i956, collected by P. F. Bellinger. One paratype has been returned to the collector. Genus SEJUS C. L. Koch Sejus Koch, C. L., (I843) 1 Ubers. d. Arachnidensyst., Heft 3 : 92. Cheiroseius Berlese, A., (1916). Redia, 12 : 33. Episeius Hull, J. E., (1918). Trans, nat. Hist. Soc. Northumb. (n.s.), 5 : 63. Episeiella Willmann, C., (1938). Ann. Hist. nat. Mus. Hung., 31 : 167. Platyseiine mites with the " posterior dorsal shield " bearing fifteen pairs of setae of which five pairs belong to the J series ; S series also comprising five pairs. Dorsal setae usually simple, rarely pilose. Lateral margins of the dorsal shield entire or incised. Sternal shield in the female with three pairs of setae ; metasternals free. Genital shield wedge shaped and bearing a pair of setae. Ventri-anal shield large with two or more pairs of pre-anals. Peritreme with or without poststigmatic process. Leg I rarely without ambulacrum, usually with at least two claws. Legs II-IV with normal ambulacrum. This is, at present, the largest genus of the subfamily. A number of the older species are inadequately described and must remain species incertae sedis until the type material can be re-examined. The following key includes only those species we have examined. Type species Sejus viduus C. L. Koch, 1839 1 PP- 73~ I 3 2 f the third Heft were published in 1843 although most authors consider the complete Heft to have appeared in 1842. 50 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) KEY TO THE SPECIES IN THE BRITISH MUSEUM (NATURAL HISTORY) FEMALES 1. Dorsal shield with lateral incisions (Text-figs. 114 and 168) .... 2 Lateral margins of the dorsal shield entire ....... 3 2. Lateral incisions located anterior to setae Si and coinciding with their position in the deutonymph (Text-fig. 114) ; median process of tectum long, Y-shaped (Text-fig. 117) ; vertical setae absent; ventri-anal shield with one pair of preanals (Text-fig. 116). Sikkim .... S. clayi sp. nov. (p. 73) Lateral incisions wide and situated between 82 and 83 (Text. -fig 168) ; median process of tectum not Y-shaped ; vertical setae present ; ventrianal shield with two pairs of preanals (Text-fig. 169). Argentina . S. ornatus sp. nov. (p. 88) 3. Tarsus I without ambulacrum (Text-fig. 12) . . . . . . 4 Tarsus I with ambulacrum or at least with two distinct claws ... 6 4. Vertical setae considerably longer than 12 (Text-fig. 53) ; poststigmatic process of the peritreme extending posterior to coxa IV (Text-fig. 54). Europe S. viduus C. L. Koch. (p. 52) Vertical setae shorter than i2 (Text-fig. 162) ; poststigmatic process very short or absent ............. 5 5. With a short but conspicuous poststimatic process of the peritreme ; sternal shield with an elongate reticulated area (Text-fig. 164). Sikkim and Argentina S. browningi sp. nov. (p. 86) Without a poststigmatic process ; sternal shield without such ornamentation (Text-fig. 144). Jamaica ...... S. hulli sp. nov. (p. 81) 6. Peritreme not produced posterior to the stigma (Text-fig. 81) or poststigmatic process narrow, short, never extending posterior to coxa IV (Text-fig. 154) . 7 Poststigmatic portion of the peritreme strongly developed and extending posterior to coxa IV (Text-fig. 56) .......... 17 7. Dorsal shield strongly arched (Text-fig. 79) ; sternal region not definitely sclero- tized (Text-fig. 81) ; a number of dorsal setae broad, pilose ; ventrianal shield with 7 setae. Britain ..... S. dromadis sp. nov. (p. 62) Dorsal shield at the most with a median longitudinal ridge ; sternal shield well- sclerotized ; dorsal setae simple or lanceolate ...... 8 8. Leg IV greatly elongated, more than twice the length of the dorsal shield (Text-fig. 99) ; vertical setae prominent, directed anteriorly ; ventri-anal shield with 9 setae. E. Africa ...... S. phalangioides sp. nov. (p. 66) Leg IV never more than i times the length of the dorsal shield ... 9 9. Vertical setae apparently absent (Text-fig. 124) ; dorsal and ventral shields charac- teristically ornamented (Text-figs. 124, 125, 129). Jamaica S. aciculatus sp. nov. (p. 76) Vertical setae present, prominent, directed anteriorly (Text-fig. 153) . . 10 10. Peritreme with a distinct poststigmatic process (Text-fig. 154) . . . . n Peritreme without a poststigmatic process (Text-fig. 141) . . . . 13 11. Dorsal shield with a distinct " tubercle " in the region of setae J4 this indicates the highest point of the median longitudinal ridge (Text-fig. 153) ; postigmatic portion of peritreme extending to about the middle of coxa IV. Jamaica 5. neborealis sp. nov. (p. 84) Dorsal shield without tubercle in the region of setae J4 . . . . . 12 12. Vertical setae shorter than J2 (Text-fig. 103) ; femur, genu and tibia of leg with sharply pointed spine-like setae situated on tubercles along their anterior and posterior margins; ventri-anal shield with 9 setae (Text-fig. 104). E. Africa 5. spinipes sp. nov. (p. 69) Vertical setae long, about equal in length to J2 (Text-fig. 63) ; chaetotaxy of leg I normal ; ventri-anal shield with n setae. Europe 5. borealis (Berl.) (p. 56) THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 51 13. Tarsus I with a long, strong seta situated dorsally towards the middle of the segment (Text-fig. 140) . . . . . . . . . . I ^ Tarsus I without such a prominent seta dorsally . . . . . . 15 14. Dorsal shield with a conspicuous tubercle overhanging setae J5 (Text-fig. 158). Jamaica . 5. nodosus sp. nov. (p. 85) Dorsal shield without a tubercle in the region of setae J4 and J5 (Text-fig. 138). Jamaica 5. bellingeri sp. nov. (p 81). 15. Ventri-anal shield broad, gently rounded posteriorly (Text-fig. 92) ; peritrematal shield not developed posterior to coxa IV. E. Africa . S. kennedyi sp. nov. (p. 65) Ventri-anal shield markedly tapering in its posterior half (Text-figs. 121 and 132), subtriangular in outline ......... 16 1 6. Sternal shield with a distinct reticulated area medially ; pre-endopodal shields absent ; paranal setae situated on a strongly-sclerotized bar (Text-fig. 132). Jamaica ........ 5. antillanus sp. nov. (p. 78) Sternal shield without reticulated area ; pre-endopodal shields present (Text- fig. 12 1) ; paranals not on sclerotized bar. Indonesia . S. alstoni sp. nov. (p. 74) 17. Claws on leg I large, sessile (Text-fig. 14) ; tarsus I with a thick spine on its anterior margin. Europe .... 5. unguiculatus (Berl.) (p. 53) Claw on leg I normal, usually situated on a pretarsus, if sessile claws minute (Text-fig. 13) . 18 18. Surface of dorsal shield minutely punctured or tuberculated (Text-fig. 55), reticula- tion when present restricted to the lateral margins of the shield . . . 19 Surface of the dorsal shield reticulated, often scabrid, without punctures (Text- fig. 74) . 20 19. Dorsal shield minutely punctured (Text-fig. 55) ; sternal shield with an oval reticulated area (Text-fig. 56) ; ventri-anal shield with n setae. Europe S. laelaptoides (Berl.) (p. 54) Dorsal shield tuberculated (Text-fig. 174) ; sternal shield lacking a reticulated area ; ventri-anal shield with 9 setae (Text-fig. 175). Argentina S. tuberculatus sp. nov. (p. 89) 20. Claws on leg I minute, apparently sessile ; leg I long and slender, about i times the length of the dorsal shield .......... 21 Claws on leg I situated on a short but distinct pretarsus ; leg I usually considerably less than ij times the length of the dorsal shield ...... 22 2 1 . Tibia and tarsus I approximately equal in length ; tarsus I more than 200/4 in length. Jamaica ........ S. jamaicensis sp. nov. (p. 83) Tibia I distinctly longer than tarsus I ; tarsus I not more than 160/1 in length. Nepal ........ S. nepalensis sp. nov. (p. 71) 22. Vertical setae long, directed anteriorly (Text-fig. 74) ; dorsal shield heavily orna- mented ; tarsus I about twice the length of tibia I. British Isles 5. cassiteridum sp. nov. (p. 61) Vertical setae relatively shorter, directed antero-laterally or posteriorly (Text- figs. 67 and 86) ; vertex not well developed ....... 23 23. Tibia I equal in length to or longer than tarsus I. Europe S. serratus (Halbert) (p. 58) Tibia I distinctly shorter than tarsus I (i : 1.3-1.7) ...... 24 24. Ventri-anal shield with five subcircular areas arranged crescentically posterior to the paranals (Text-fig. 87) ; small weakly-sclerotized species about 360^ in length. E. Africa . S. areolatus sp. nov. (p. 64) Ventri-anal shield without such areas ; larger species 400-800/4 in length . . 25 25. J5 about J the length of J4, setiform (Text- fig. no). Nepal S. parbatensis sp. nov. (p. 72) J5 less than i the length of J4, usually thorn-like ...... 26 52 THE PLATYSEIINAE (ME SOSTIGM AT A : ACEOSEJIDAE) 26. Sternal shield without semicircular depressions near its anterior margin (Text-fig. 71). Europe ........ S. curtipes (Halbert) (p. 59) Sternal shield with a pair of semicircular depressions near its anterior margin (Text-fig. 60) 27 27. Sclerotized bases of the vertical setae large, almost contiguous (Text-fig. 59) ; highest point of median dorsal ridge located between 15. Europe 5. necorniger (Oudemans) (p. 55) Bases of vertical setae weakly sclerotized, widely separated (Text-fig. 100) ; highest point of median dorsal elevation located between i5 and Ji. E. Africa S. signatus sp. nov. (p. 68) Sejus viduus C. L. Koch Sejus viduus Koch, C. L., 1839, Deutsch. Crust. Myr. Arach., Fasc. 24, t. 10. Epicrius corniger Berlese A., 1891, A. M.S., Fasc. 59, No. 2. Sejus viduus C. L. Koch, Oudemans, A. C., 1936, K.H.O.A., 3A : 376. Episeiella heteropoda Willmann, C., 1938, Ann. Hist. nat. Mus. Hung., 31 : 164, syn. nov. FEMALE. Dorsal shield (529 ju, long x 330 p wide) covered by a strong irregular 54 FIGS. 53-54. Sejus viduus (C. L. Koch), female. Fig. 53, dorsal shield. Fig. 54, venter. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 53 network. Lateral margins irregular. "Anterior dorsal shield" with twenty-one pairs of simple setae ; verticals very prominent, well separated, and directed anteriorly. " Posterior dorsal shield " with fifteen pairs of simple setae of which J5 are the shortest. All the setae arise from small protuberances (Text-fig. 53). Tritosternum with a pair of pilose laciniae. Sternal shield with a reticulated area medially and bearing three pairs of simple setae. Genital shield slightly convex posteriorly and bearing a single pair of simple setae. Ventri-anal shield broader than long (101 /* long x 139 /* wide) with only faint traces of reticulations, and bearing nine simple setae. Between the genital and ventri-anal shields lie four narrow sclerotized platelets. Stigma situated between coxae III and IV ; posterior prolongation of peritreme, and peritrematal shield, reaching beyond the posterior margin of coxa IV (Text-fig. 54). Metapodal shields small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged, and chelicerae typical of the genus. Leg I (724 /* long) with the tarsus (144 JLI) shorter than the tibia (167 /*). Setae on tarsus, tibia and genu I very fine ; tarsal claws absent ; remaining setae stouter. Tarsi II-IV with a pair of lanceolate setae ; ambulacra without acuminate lobes. MALE. Unknown. DISTRIBUTION. Koch's (1839) description is based on material from wet woodland habitats in Germany. Berlese's (1891) description of Epicrius corniger is based on several specimens in moss on high mountains, Venice. Berlese's type has been examined by us in Florence. Willmann's (1939) description of Episeiella heteropoda, was made from one female sifted at Vihnyei volgy, Hungary, and is the specimen from which our description and figures are made. Sejus unguiculatus Berlese Sejus unguiculatus Berlese, A., 1887, A. M.S., 41, No. 4. Lasioseius (Cheiroseius) unguiculatus Berlese, A., 1916, Redia, 12 : 33. The characteristic feature of this species is that the claws on tarsus I are greatly enlarged and sessile (Text-fig. 14). The chaetotaxy of the dorsal shield is typical of the genus. The " posterior dorsal shield " bears fifteen pairs of setae ; five pairs in the J series. The " anterior dorsal shield " bears twenty pairs of setae ; vertex weakly developed. The chaetotaxy and sclerotization of the venter are typical of Sejus. Sternal shield with three pairs of setae ; genital shield with one pair of setae ; ventri-anal shield large, occupying most of the area posterior to coxae IV, and bearing eleven simple setae. Six small plates are situated between the genital and ventri-anal shields. Peritreme extending posterior to coxa IV. Metapodal plates conspicuous. Berlese described this species from moss in Venice. The above description is based on drawings of the type. This species is not represented in the Museum Collections. 54 THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) Sejus Ixlaptoides (Berlese) comb. nov. Epicrius laelaptoides Berlese, A., 1887, A. M.S., Fasc. 40, No. 10. Parasejus glaber var. minor Tragardh, I., 1910, Naturw. Untersuch. Sarekgeb., 4 : 432, syn. nov. Lasioseius (Episeius) sphagni Halbert, J. N., 1923, /. Linn. Soc. Lond. (Zool.), 35 : 371, syn. nov. FEMALE. Dorsal shield (518-581 /* long x 330-415 ju, wide) minutely punctured and with a few faint scattered reticulated areas (Text-fig. 55). " Anterior dorsal shield ' ' with twenty-one pairs of simple setae ; verticals well separated and directed anteriorly. " Posterior dorsal shield " with fifteen pairs of simple setae ; J5 very 56 FIGS. 55-58. Sejus laelaptoides (Berlese), female. Fig. 55, dorsum. Fig. 56, venter. Fig. 57, chelicera. Fig. 58, tarsus II. short. The distribution and relative lengths of the dorsal setae, and the ornamenta- tion of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with a con- spicuous elliptical reticulated area, and bearing three pairs of simple setae (Text- fig. 56). Genital shield broad, posterior margin straight, and bearing one pair of simple setae. Ventri-anal shield large, broader than long (152-215 /i long x 240- 291 fi wide), reticulated, and bearing eleven simple setae. Between the genital and ventri-anal shields lies a row of four narrow sclerotized plates. Stigma situated between coxae III and IV ; peritreme and peritrematal shield extending posterior to coxa IV. Metapodal plates small. THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 55 Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 57. Leg I (536-673 IJL long) with the tarsus (113-126 f) longer than the tibia (104- 118/4). Setae on tarsus and tibia fine, those on other segments somewhat stouter ; tarsal claws strong. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 58). MALE. Unknown. DISTRIBUTION. Berlese (1887) described Epicrius laelaptoides from the Royal Botanic Gardens, Padua. Tragardh (1910) described Parasejus glaber var. minor from Sphagmim at Sarekgebirge, Sweden, and Halbert (1923) described Lasioseius (Episeius) sphagni from Sphagnum in Co. Dublin, Eire. The writers have examined Berlese's type in Florence, Halbert's type, and also numbers of females from Whern- side Moor, Cock Hill, and Malham Tarn, Yorkshire ; Rydal Water and Grasmere, Westmorland ; Leith Hill and Beckley, Surrey ; Lapland ; and Michigan. Sejus necorniger (Oudemans) comb. nov. Hypoaspis necorniger Oudemans, A. C., 1903, Ent. Ber., 1, 12 : 87. Platyseius necorniger, Buitendijk, A. M., 1945, Zool. Med., 24 : 296. Episeius necorniger, Willmann, C., 1949, Veroff. Mus. Nat. Bremen, No. lA : 120. Episeius necorniger, Willmann, C., 1952, Veroff. Inst. Meeresf. Bremerhaven, 1 : 148. FEMALE. Dorsal shield (550-722 /* long x 320-467 fi wide) completely covered by a fine reticulated pattern (Text-fig. 59). " Anterior dorsal shield " with twenty- one pairs of simple setae ; verticals, with their bases close together, considerably shorter than post- verticals. " Posterior dorsal shield " with fifteen pairs of simple setae. The distribution of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with some sculpturing, and with two semicircular marks anteriorly, and bearing three pairs of simple setae (Text-fig. 60). Genital shield convex posteriorly and bearing a single pair of setae. Ventri-anal shield slightly wider than long (162-190 [i long X 177- 228 fi wide), reticulated, and bearing nine simple setae. Between the genital and ventri-anal shields lie four small sclerotized plates. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates narrow. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 61. Leg I (578-782 IJL long) with the tarsus (144-184 /*) longer than the tibia (101- 126 fi) ; claws small, pulvillus short. Setae on tarsus and tibia fine, those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 62). MALE. Unknown. 56 THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) DISTRIBUTION. Oudemans (1903) did not state the locality for his type female, but Buitendijk (1945) in his catalogue of the Oudemans Collection gives the locality as Borkum Island, Germany. Willmann (1952) records the species from the North Sea island of Wangerooge, and says it is found as far up as Lapland. We have examined specimens from the following localities : Kirkstone Pass, River Brathey and High Pike, Ambleside, Westmorland ; Rambulls Moor, Ilkley, Yorkshire ; FIGS. 59-62. Sejus necorniger (Oudemans), female. Fig. 59, dorsum. Fig. 60, venter. Fig. 61, chelicera. Fig. 62, tarsus II. Hathersage, Derbyshire ; Sea Houses, Northumberland ; Burton Marsh, Cheshire ; Llyn Idwal, Caernarvonshire ; South Wales ; Washford, Somerset ; Kiel, Germany ; Ohrid, Montenegro ; and Grabouw, Transvaal. Sejus borealis (Berlese) comb. nov. Ameroseius borealis Berlese, A., 1904, Redia, 1 : 259. Episeius montanus Willmann, C., 1949, Veroff. Mus. Nat. Bremen, No. lA : 120. syn. nov. Episeius montanus Willmann, C., 1952, Veroff. Inst. Meeresf. Bremerhaven, 1 : 148. Episeius montanus Willmann, C., 1952, SB. ost. Akad. Wiss., Abt. i, 162 : 462. FEMALE. Dorsal shield (570-581 JJL long x 361-371 fi wide) covered by a reticu- lated pattern (Text-fig. 63). " Anterior dorsal shield " bears twenty-one pairs of simple setae; verticals anteriorly projected. "Posterior dorsal shield" bears THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) 57 fifteen pairs of simple setae ; J5 are at the most half the length of J4. The majority of the dorsal setae arise from small tubercles. The distribution and relative length of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed and flanked at its base by a pair of narrow pre- endopodal shields. Sternal shield with a narrow longitudinal reticulated area, and bearing three pairs of simple setae (Text-fig. 64). Genital shield convex posteriorly and bearing a single pair of setae. Ventri-anal shield broader than long (139-164 [i long X 202-205 fi wide), reticulated and bearing eleven simple setae. Between the genital and ventri-anal shield lies a row of four small plates, and, posterior to these, FIGS. 63-66. Sejus borealis (Berlese), female. Fig. 63, dorsum. Fig. 64, venter. Fig. 65, chelicera. Fig. 66, tarsus II. two smaller plates. Stigma situated between coxae III and IV with the posterior prolongation of the peritreme extending to about the middle of coxa IV. Metapodal plates circular. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 65. Leg I (576-580 fi long) with the tarsus (137-139 /*) longer than the tibia (108- 114/4). Setae on tarsus and tibia I fine ; ambulacrum and claws small ; remaining setae stouter and arising from small protuberances. Tarsi II-IV with a pair of ZOOL. 6, NO. 2. 5 58 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) lanceolate setae, and ambulacra with three acuminate lobes (Text-fig. 66). Setae of legs II-IV arising mainly from small protuberances. MALE. Unknown. DISTRIBUTION. Berlese's (1904) description of the female of Ameroseius borealis is based on material collected by Thor in Norway. His type specimen has been examined by us. Willmann (1949) described Episeius montanus from Poland ; in 1952 he recorded the species from the North Sea island of Wangerooge, and in 1953 he recorded it from several localities in the eastern Alps. The writers have examined five females from Soil Insecticide Experiment, Bellahouston Park, Glasgow, October, 1951, collected by J. G. Sheals ; one female from ditch-side debris near Markfield, Leicestershire, 19.^.1958, collected by P. N. Lawrence ; and one female from wet moss at West Carter Basin, Labrador, I7.viii.i958, also collected by P. N. Lawrence. Sejus serratus (Halbert) comb. nov. Paraseius serratus Halbert, J. N., 1915, Proc. R. Irish Acad,, 39ii : 78. Lasioseius (L.) serratus, Schweizer, J., 1922, Ver. naturf. Ges. Basel, 33 : 43. Episeius serratus, Schweizer, J., 1949, Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 63. Episeius serratus, Willmann, C., 1949, Veroff. Mus. Bremen, 1A : 120 ; 1954, I n Spiegel, Land-Tierwelt, 15 Ordnung, Acarina .'351. FEMALE. Dorsal shield (440-570 fi long x 286-351 /i wide) with a pattern of irregular reticulations and depressions (Text-fig. 67). "Anterior dorsal shield" with twenty-one pairs of simple setae ; verticals on small tubercles and directed laterally. "Posterior dorsal shield" with fifteen pairs of simple setae. The distribution of the dorsal setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum with a narrow base and pilose laciniae. Sternal shield with some sculpturing laterally, with a pair of circular marks anteriorly, and bearing three pairs of simple setae (Text-fig. 68). Genital shield convex posteriorly and bearing a single pair of setae. Ventri-anal shield slightly wider than long (152-170 /* long X 165-200 /* wide), reticulated anteriorly and bearing nine simple setae. Between the genital and ventri-anal shields lie a row of small plates and behind these a further pair. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 69. Leg I (528-718 /i long) with the tarsus (121-149 /) shorter than the tibia (139- 164 fi). Claws on leg I small ; pulvillus short. Setae on tarsus and tibia fine ; those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes. MALE. This sex is described and figured by Schweizer (1922 & 1949). DISTRIBUTION. Halbert (1915), in the Clare Island Survey, described this species from specimens " Found commonly in sphagnum, gathered on the slopes of Croagh- THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 59 patrick at an elevation of about 600 ft., during the month of October ". Schweizer (1922 & 1949) records both sexes from the Swiss Alps, and Willmann (1954) records the female from Austria. The Museum Collections contain the following specimens : one female in Sphagnum from marsh, Raise Beck, Grasmere, Westmorland, 900 ft., 29. vi. 1951, collected by J. T. Salmon ; three females in mosses and liverworts on rocks by south-east bank of Rydal Water, Westmorland, 29. vi. 1956, collected by FIGS. 67-69. Sejus serratus (Halbert), female. Fig. 67, dorsum. Fig. 68, venter. Fig. 69, chelicera. P. N. Lawrence ; one female in Sphagnum at Blelham Tarn, Lancashire, 22 .iii. 1955, collected by R. J. Elliott ; one female in Sphagnum at Box Hill, Surrey, 30. vi. 1951, collected by E. Duffey ; and two females in moss at West Carter Basin, Labrador, 7.viii.i958, collected by P. N. Lawrence. Sejus curtipes (Halbert) comb. nov. Lasioseius (Episeius) glaber var. curtipes Halbert, J. N., 1923., /. Linn. Soc. Lond. (Zool.), 35 : 37- Episeius ovaspini Schweizer, J., 1949, Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 69, syn. nov. FEMALE. Dorsal shield (456-529 /* long x 258-320 fi wide) with a reticulated pattern (Text-fig. 70). " Anterior dorsal shield " with fifteen pairs of simple setae. 6o THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) The distribution of the setae, and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum with pilose laciniae. Sternal shield very faintly sculptured and bearing three pairs of simple setae (Text-fig. 71). Genital shield convex posteriorly and bearing a single pair of setae. Ventri-anal shield wider than long (120-160 /* long X 154-218 11 wide), reticulated in its anterior half, and bearing nine simple setae. Between the genital and ventri-anal shields lies a pair of long narrow platelets and behind these two shorter platelets. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates small. FIGS. 70-73. Sejus curtipes (Halbert), female. Fig. 70, dorsum. Fig. 71, venter. Fig. 72, chelicera. Fig. 73, tarsus II. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 72. Leg I (443-518 fi long) with the tarsus (113-139 /i) longer than the tibia (68- 88 fi). Claws on leg I small, pulvillus short ; setae on tarsus and tibia fine, those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 73). MALE. This sex is described and figured by Schweizer (1949 : 71). DISTRIBUTION. Halbert's (1923) description of Lasioseius (Episeius) glaber van THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 61 curtipes is based on two females found amongst wet moss on Lambay Island, Co. Dublin, in July. Schweizer's (1949) description of Episeius ovaspini is based on five males and thirty-two females collected in wet mosses in the Swiss National Park. We have examined Halbert's type and a number of Schweizer's type series ; also five females from the summit of Snowdon, Caernarvonshire, August, 1951, collected by E. Duffey ; and eleven females from Jamaica, B.W.I., collected by P. F. Bellinger ; seven in moss on concrete, Coolsbade, Portland, I7.vii.i955, and four in moss, west side of John Cove Mountains, 2,300 ft., 25. v. 1956. Sejus cassiteridum sp. nov. FEMALE. Dorsal shield (435-467 fj, long x 237-258 /* wide) heavily sculptured and with a network of ridges and depressions. Lateral margins irregular (Text-fig. 74). "Anterior dorsal shield" with twenty-one pairs of simple setae; verticals 75 FIGS. 74-78. Sejus cassiteridum sp. nov., female. Fig. 74, dorsum. Fig. 75, venter. Fig. 76, tectum. Fig. 77, chelicera. Fig. 78, tarsus II. close together and projecting anteriorly. "Posterior dorsal shield" with fifteen pairs of simple setae of which J5 are the shortest. The distribution and relative lengths of the dorsal setae and the ornamentation of the dorsal shield are shown in the Text-fig. 62 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Tritosternum well developed with pilose laciniae. Sternal shield almost entirely without sculpturing, and bearing three pairs of short simple setae (Text-fig. 75). Genital shield slightly convex posteriorly and bearing a single pair of setae. Ventri- anal shield wider than long (115-139 fi long X 152-165 /* wide), reticulated and bearing nine simple setae. Between the genital and ventri-anal shields lie six small sclerotized plates. Stigma situated between coxae III and IV ; peritreme and peritrematal shields extending posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 76) three-pronged. The dentition of the chelicera is shown in Text-fig. 77. Leg I (372-387 fi long) with the tarsus (99-106 fi) almost twice the length of the tibia (58-63 fi). Setae on tarsus and tibia fine, those on other segments stouter ; pulvillus short and tarsal claws small. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 78). MALE. Unknown. LOCALITY. England. The holotype female (1959.1.20.92) and twenty-five paratype females (1959 . i . 20 . 93-102) occurring with Platyseius subglaber (Oudemans) in the roots of rushes by Big Pool, St. Agnes, Isles of Scilly, Cornwall, 19. ix. 1957, collected by K. H. Hyatt. The writers have examined a single female (1959.1.20.193) from Sphagnum, Blelham Tarn, Windermere, Lancashire, 22 .iii. 1955, collected by R. J. Elliott, which differs from the St. Agnes specimens by having the sternal shield conspicuously reticulated. The measurements of this specimen all fall within the ranges given above for Sejus cassiteridum (viz. : dorsal shield 446 x 247 fi, ventri-anal shield 126 x 157 11, leg I 392 /*, tarsus I 101 /*, tibia I 63 ju,). Sejus dromadis sp. nov. FEMALE. Dorsal shield (393 /* long x 230 p wide) oval in outline, heavily reticulated, and steeply humped dorso-medially (Text-fig. 79). " Anterior dorsal shield " with twenty-three pairs of setae ; " posterior dorsal shield " with fifteen pairs of setae (Text-fig. 80). The verticals are stout and spiculate, and the re- mainder of the setae are stout and pilose. The distribution and relative lengths of the dorsal setae, and the form of the dorsal shield are shown in the Text-figs. Tritosternum less strongly developed than in the majority of species of Sejus. Laciniae pilose. Sternal shield weakly sclerotized with three pairs of simple setae. Genital shield broadly flask-shaped with one pair of setae (Text-fig. 81). Ventri- anal shield small (83 fi long x 88 ju, wide), weakly sclerotized, and bearing seven simple setae. Two small sclerotized plates lie on the interscutal membrane between the genital and ventri-anal shields. Stigma situated between coxae III and IV, with the peritreme not extending posterior to the stigma. Peritrematal shield extending a little posterior to the stigma. Venter of gnathosoma with rostral and internal palptrochanter setae long and whip-like (Text-fig. 82). All setae on pedipalp simple. Tectum three-pronged and THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 82 FIGS. 79-85. Sejus dromadis sp. nov., female. Fig. 79, lateral view of dorsum. Fig. 80, dorsal shield. Fig. 81, venter. Fig. 82, venter of gnathosoma. Fig. 83, tectum. Fig. 84, chelicera. Fig. 85, tarsus II. 6 4 THE PLATYSEIINAE (MESOSTI GM ATA : ACEOSEJIDAE) typical of the genus (Text-fig. 83). The dentition of the chelicera is shown in Text-fig. 84. Leg I (about 300 /* long) with the tarsus (70 /*) longer than the tibia (56 fi). Setae on tarsus fine ; those on remaining segments stouter, and in some cases spiculate and arising from small protuberances. Tarsi II-IV with a pair of lanceolate setae. Ambulacrum of tarsus I short and claws small. Pulvilli of tarsi II-IV produced into three acuminate lobes (Text-fig. 85). MALE. Unknown. LOCALITY. England. A single female (1959.1.20.103) in cow dung at Canter- bury, Kent, collected by the late Dr. E. Warren. Sejus areolatus sp. nov. FEMALE. Dorsal shield (361 fi long X 216 ju, wide) completely covered by a fine reticulated pattern; margin irregular (Text-fig. 86). "Anterior dorsal shield" FIGS. 86-90. Sejus areolatus sp. nov., female. Fig. 86, dorsum. Fig. 87, venter. Fig. tectum. Fig. 89, chelicera. Fig. 90, tarsus II. with twenty-one pairs of simple setae ; verticals with their bases touching, and considerably shorter than post-verticals. "Posterior dorsal shield" with fifteen pairs of simple setae. The distribution of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield almost plain, THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 65 with three pairs of simple setae (Text-fig. 87). Genital shield bearing a single pair of setae. Ventri-anal shield wider than long (106 /* long x 132 /* wide), reticulated, with five small oval areas forming a crescent behind the paranals, and bearing nine simple setae. Between the genital and ventri-anal shields lies a row of four sclero- tized platelets, and behind these two smaller ones. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates small and narrow. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged (Text-fig. 88), the centre prong being in advance of the lateral ones. The dentition of the chelicera is shown in Text-fig. 89. Leg I (c. 330 ju, long) with the tarsus (88 /*) longer than the tibia (50 /*). Setae on tarsus and tibia fine, those on remaining segments stouter ; claws small, on short pulvillus. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 90). MALE. Unknown. LOCALITY. Uganda. A single female, the holotype (1959.1.20.104), from rushes and grassland bordering hot springs, sixteen miles east-north-east of Bundi- bugyo, Ruwenzori, 29.viii.i952, collected by G. Owen Evans. Sejus kennedyi sp. nov. FEMALE. Dorsal shield (456-467 jn long x 268-279 /i wide) covered by a network of small depressions and ridges (Text-fig. 91). " Anterior dorsal shield " with twenty-one pairs of simple setae ; verticals separated by the width of their bases and projecting anteriorly. "Posterior dorsal shield" with fifteen pairs of simple setae of which J5 are short. The distribution and relative lengths of the dorsal setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum with a pair of pilose laciniae. Sternal shield plain and bearing three pairs of simple setae (Text-fig. 92). Genital shield lightly reticulated, convex posteriorly, and bearing one pair of simple setae. Ventri-anal shield broader than long (139-144 ju, long x 208-213 fi wide), reticulated, and bearing nine simple setae. Between the genital and ventri-anal shields lie two narrow sclerotized platelets. Stigma situated between coxae III and IV ; peritreme crenate along its internal margin, and with no post-stigmal prolongation. Peritrematal shield tapering and extending posterior to coxa IV. Metapodal shields obliquely situated. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. Chelicerae typical. Leg I (354 /* long) with the tarsus (88 /*) longer than the tibia (63 /*). Setae on tarsus I fine, ambulacrum and claws small ; setae on remaining segments stouter and arising from small protuberances. Tarsi II-IV with a pair of lanceolate setae and ambulacra with three acuminate lobes. MALE. Unknown. 66 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) LOCALITY. Uganda. The holotype female (1959.1.20.105) and three paratype females (1959.1.20.106-108) from litter under hardwood stand, Bimdibugyo, Ruwenzori, 3,400 ft., 2.ix.i952, collected by G. Owen Evans. This species is named after Professor W. Q. Kennedy, University of Leeds, leader of the British Ruwenzori Expedition, 1952. 91 92 FIGS. 91-92. Sejus kennedyi sp. nov., female. Fig. 91, dorsum. Fig. 92, venter. Sejus phalangioides sp. nov. FEMALE. Dorsal shield (398-410 fi long x 250-253 // wide) attenuated in the posterior half, and covered by a fine reticulated pattern (Text-fig. 93). " Anterior dorsal shield " with twenty-one pairs of simple setae, verticals projecting anteriorly ; "posterior dorsal shield" with fifteen pairs of simple setae, of which J5 are the shortest. The distribution and relative lengths of the dorsal setae are shown in the Text-fig. Tritosternum well developed and flanked at its base by a pair of narrow pre- endopodal shields. Sternal shield with a reticulated area anteriorly, and bearing three pairs of simple setae (Text-fig. 94). Genital shield broad, truncated posteriorly, THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 67 distinctly reticulated, and bearing a single pair of setae. Ventri-anal shield slightly broader than long (101 fi long X 116 fi wide), attenuated posteriorly, reticulated, and bearing nine simple setae. Between the genital and ventri-anal shields lies a narrow, undulated, sclerotized strip of chitin, and two platelets. Stigma situated between coxae III and IV, peritreme scarcely extending posterior to the stigma. FIGS. 93-99. Sejus phalangioides sp. nov., female. Fig. 93, dorsum. Fig. 94, venter. Fig. 95, tectum. Fig. 96, chelicera. Fig. 97, leg IV. Fig. 98, tarsus II. Fig. 99, entire mite, showing relative lengths of legs and idiosoma. Peritrematal shield fused with well developed podal plates, which extend posterior to coxa IV. Metapodal plates small, situated close to podal plates. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- 68 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) like. All setae on pedipalp simple. Tectum (Text-fig. 95) three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 96. Leg I (536 [i long) has the tarsus (139 fi) longer than the tibia (100-106 /*). Setae on tarsus I very fine ; those on remaining segments stouter. Ambulacrum of tarsus I short, claws small. Leg IV (954 fji long) more than twice the length of the idiosoma, with the tarsus about 400 ju, long (Text-fig. 97). Tarsi II-III with a pair of lanceolate setae, and pulvilli produced into three acuminate lobes (Text-fig. 98). Tarsus IV without lanceolate setae, and pretarsus long (55 /*). The relative lengths of the legs and the idiosoma are shown in Text-fig. 99. MALE. Unknown. LOCALITY. Uganda. The holotype female (1959.1.20.109) and one paratype female (1959.1.20.110) in grassland in forest clearing near stream, nine miles north-east of Bundibugyo, Ruwenzori, 2,850 ft., 24.viii.i952, collected by G. Owen Evans. Sejus signatus sp. nov. FEMALE. Dorsal shield (581 ju, long x 371 fi wide) weakly sclerotized and covered by a reticulated pattern (Text-fig. 100). " Anterior dorsal shield " with twenty FIGS. 100-102. Sejus signatus sp. nov., female. Fig. 100, dorsum. Fig. 101, venter. Fig. 1 02, chelicera. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 69 pairs of simple setae ; verticals considerably shorter than post-verticals, and well separated. " Posterior dorsal shield " with fifteen pairs of simple setae. The distribution of the dorsal setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum with pilose laciniae. Sternal shield weakly sclerotized, with an M-shaped mark anteriorly and bearing three pairs of simple setae (Text-fig. 101). Genital shield weakly sclerotized and bearing one pair of setae. Ventri-anal shield wider than long (152 /* long x 197 /* wide), weakly sclerotized and bearing nine simple setae. Between the genital and ventri-anal shields lie a narrow sclerotized strip and two platelets. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates very small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 102. Leg I (573 fi long) with the tarsus (141 fi) longer than the tibia (99 /*). Claws on leg I small, inconspicuous, pulvillus short. Setae on tarsus and tibia fine, those on other segments stouter. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes. MALE. Unknown. LOCALITY. Uganda. A single female, the holotype (1959.1.20.111), from damp decaying vegetation and soil under elephant-grass at Ibanda, Ruwenzori, 26 . vii . 1952, 4,760 ft., collected by G. O. Evans. Sejus spinipes sp. nov. FEMALE. Dorsal shield (392 fi long X 216 /* wide) covered by a light reticulated pattern (Text-fig. 103). " Anterior dorsal shield " bears twenty-one pairs of simple setae ; verticals anteriorly projecting. " Posterior dorsal shield" bears 15 pairs of simple setae ; J5 are as long as J4- The majority of the dorsal setae arise from small tubercles. The distribution and relative lengths of the dorsal setae, and the orna- mentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with the central area reticulated, and bearing three pairs of short simple setae (Text-fig. 104). Genital shield with a single pair of simple setae. Ventri-anal shield broader than long (119 fi long X 157 /* wide), distinctly reticulated and bearing nine simple setae. Between the genital and ventri-anal shields lies a line of four contiguous sclerotized plates, and, postero-laterally to these, four smaller plates. Stigma situated between coxae III and IV, with the posterior prolongation of the peritreme extending to about the middle of coxa IV. The peritrematal shield extends posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 105) three-pronged. The dentition of the chelicera is typical of the genus. 7 o THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE Leg I ( c. 330 fi long) with the tarsus (76 fi) longer than the tibia (53 /<). Setae on tarsus I fine ; ambulacrum and claws small ; many of the setae on remaining segments stout and arising from conspicuous protuberances. Tarsi II-IV with a pair of lanceolate setae, and ambulacra with three acuminate lobes. Setae of legs II-IV mainly stout and arising from conspicuous protuberances. MALE. Unknown. LOCALITIES. Uganda and Sudan. The holotype female (1959.1.20.112) from FIGS. 103-105. Sejus spinipes sp. nov., female. Fig. 103, dorsal shield. Fig. 104, venter. Fig. 105, chelicera. rushes and grassland bordering hot springs, sixteen miles east-north-east of Bundi- bugyo, Ruwenzori, Uganda, 29.viii.i952, collected by G. O. Evans; and two paratype females (1959.1.20.113-114) on Papyrus in the Sudd Region, Southern Sudan, collected by I. W. B. Thornton. THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 71 Sejus nepalensis sp. nov. FEMALE. Dorsal shield (456-498 //. long X 289-299 fi wide) completely covered by a fine reticulated pattern (Text-fig. 106). " Anterior dorsal shield " with twenty- one pairs of simple setae ; vertical setae relatively short and directed laterally ; vertex not well developed. " Posterior dorsal shield " with fifteen pairs of simple setae (in the holotype seta Zi is missing from the left side) ; seta J5 is the shortest. All dorsal setae arise from small tubercles. The distribution and relative lengths of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed, with pilose laciniae, sternal shield with some 106 FIGS. 106-109. Sejus nepalensis sp. nov., female. Fig. 106, dorsum. Fig. 107, venter. Fig. 108, chelicera. Fig. 109, tarsus II. sculpturing, a pair of circular areas anteriorly, and bearing three pairs of simple setae (Text-fig. 107). Genital shield almost parallel sided and bearing one pair of simple setae. Ventri-anal shield (126 /* long X 124-127 //. wide) almost circular in outline, bearing nine simple setae ; anus more or less central. Between the genital and ventri-anal shields lie about four narrow platelets. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates very small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 108. 72 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Leg I (630-769 fi long) with the tarsus (126-157 /*) shorter than the tibia (150- 195 fi}. Setae on tarsus, tibia and genu very fine ; those on remaining segments stouter. Claws very small, pulvillus short. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 109). MALE. Unknown. LOCALITY. Nepal. The holotype female (1959.1.20.115) from sandy turf on open grazed hillside facing south at Siklis (28 22' N., 84 6' E.), 7,000 ft., 20- 21. iv. 1954; and one paratype female (1959.1.20.116) from damp mossy and grassy earth under shade of trees and rocks at Bakhri Kharka (28 22-5' N., 84 7-5' E.), 5,500 ft., 24. iv. 1954, collected by K. H. Hyatt. Sejus parbatensis sp. nov. FEMALE. Dorsal shield (402-435 /* long x 237-268 ^ wide) covered by an irregular reticulated pattern (Text-fig, no). " Anterior dorsal shield " with twenty- FIGS. 110-113. Sejus parbatensis sp. nov., female. Fig. no, dorsum. Fig. in, venter. Fig. 112, chelicera. Fig. 113, tarsus II. one pairs of simple setae ; verticals with their bases contiguous. " Posterior dorsal shield " with fifteen pairs of simple setae ; setae J5 are more than half the length of J4. All the setae arise from small tubercles. The distribution of the setae, and the ornamentation of the dorsal shield are shown in the Text-fig. THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 73 Tritosternum well developed and with pilose laciniae. Sternal shield lightly sculptured and bearing three pairs of simple setae (Text-fig, in). Genital shield bearing a single pair of setae. Ventri-anal shield wider than long (114-126 ju, long X 157-170 /* wide), reticulated, and bearing nine simple setae. Between the genital and ventri-anal shields lies a row of four narrow platelets, and, posterior to these, two smaller ones. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates very small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 112. Leg I (c. 340 /* long) with the tarsus (80-83 /*) longer than the tibia (50-53 ju). Setae on tarsus I very fine ; those on remaining segments stouter ; claws small, on short pulvillus. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 113). MALE. Unknown. LOCALITY. Nepal. The holotype female (1959.1.20.117) and one paratype (1959.1.20.118) from litter under a group of deciduous trees forming a canopy on open ground at Gurjakhani (28 36-5' N., 83 13-5' E.) on the southern slope of the Dhaulagiri Himal, 8,500 ft., 24-27^1.1954 ; and one paratype female (1959.1.20. 119) in sandy turf on open, grazed hillside facing south at Siklis (28 22' N., 84 6' E.), 7,000 ft., 2O-2i.iv. 1 954, collected by K. H. Hyatt. Sejus clayi sp. nov. FEMALE. Dorsal shield (661 /* long X 366 /* wide) with lateral incisions anterior to Si, and heavily marked with circular depressions (Text-fig. 114). Vertex of dorsal shield produced anteriorly (Text-fig. 115). "Anterior dorsal shield" with twenty pairs of setae ; verticals absent, para-verticals short and spatulate, remaining setae simple. " Posterior dorsal shield " with fifteen pairs of simple setae. Setae J5 are long. The distribution and relative lengths of the dorsal setae are shown in Text-fig. 114. Tritosternum well developed with pilose laciniae and flanked at its base by a number of fragmented pre-endopodal shields. Sternal shield plain and bearing three pairs of simple setae (Text-fig. 116). Genital shield wedge shaped, genital setae off the shield. Ventri-anal shield slightly wider than long (185 fi long X 190 /* wide), covered with small circular depressions, and bearing only five simple setae. Stigma situated between coxae III and IV, peritreme not extending posterior to the stigma; external margin of peritreme strongly crenate. Peritrematal shield extending posterior to coxa IV. Metapodal shields narrow and obliquely situated. Venter of the gnathosoma with rostral and internal palptrochanter setae long, whip-like. All setae on legs simple. Tectum (Text-fig. 117) differing from that of other members of the genus in that the median prong is strongly produced and terminates in a pair of pilose laciniae. The dentition of the chelicera is shown in Text-fig. 118. ZOOL. 6, NO. 2. 74 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Leg I (481 /* long) with the tarsus (126 /*) longer than the tibia (101 /i). Setae on tarsus I very fine ; those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae and ambulacra with lateral lobes acuminate, median lobe short and rounded (Text-fig. 119). Ambulacrum of tarsus I short with claws very small. MALE. Unknown. 115 117 \ I) 119 118 FIGS. 114-119. Sejus clayi sp. nov., female. Fig. 114, dorsum. Fig. 115, vertex. Fig. 116, venter. Fig. 117, tectum. Fig. 118, chelicera. Fig. 119, tarsus II. LOCALITY. Sikkim. A single female (1959.1.20.120) from leaves and leaf- mould in wood on steep hillside, Chungtang, 5,120 ft., 16.^.1952, collected by Dr. Theresa Clay. Sejus alstoni sp. nov. FEMALE. Dorsal shield (507-550 ju, long X 279-340 /* wide) covered anteriorly by a network of small depressions and ridges, and posteriorly by a reticulated THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 75 pattern (Text-fig. 120). " Anterior dorsal shield " with twenty-one pairs of simple setae ; verticals close together and directed anteriorly. " Posterior dorsal shield " with fifteen pairs of simple setae, of which J5 are the shortest. The distribution and relative lengths of the dorsal setae, and the ornamentation of the dorsal shield are shown in the Text-fig. FIGS. 120-123. Sejus alstoni sp. nov., female. Fig. 120, dorsum. Fig. 122, tectum. Fig. 123, tarsus II. 121 Fig. 121, venter. Tritosternum well developed with pilose laciniae, and flanked at its base by a pair of narrow pre-endopodal shields. Sternal shield weakly sclerotized, without ornamentation, and bearing three pairs of simple setae (Text-fig. 121). Genital shield truncated posteriorly and bearing one pair of setae. Ventri-anal shield broader than long (139-152 IJL long X 180-205 / wide), reticulated, and bearing nine simple setae. Between the genital and ventri-anal shields lies a row of three narrow sclerotized plates. Stigma situated between coxae III and IV ; no posterior prolongation of the peritreme. Peritrematal shield extending posterior to coxa IV. Metapodal shields elliptical. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- 76 THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) like. All setae on pedipalp simple. Tectum (Text-fig. 122) three-pronged and typical of the genus. Chelicerae typical. Leg I (560-565 fi long) with the tarsus (160 p) one-and-a-half times as long as the tibia (101-104 /*). Setae on tarsus fine ; claws small ; setae on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae and ambulacra with three acuminate lobes (Text-fig. 123). MALE. Unknown. LOCALITY. Indonesia. The holotype female (1959.1.20.121) and two paratype females (1959.1.20.122-123) in decaying vegetation at Bogor, 1954, collected by the late A. H. G. Alston. Sejus aciculatus sp. nov. FEMALE. Dorsal shield (329-341 /* long x 223-231 p wide) very heavily sclero- tized, entirely covered by a network of ridges and depressions (Text-fig. 124). " Anterior dorsal shield" with nineteen pairs of simple setae, verticals and para- verticals apparently absent. " Posterior dorsal shield " with fifteen pairs of simple setae. The distribution and relative lengths of the dorsal setae, and the charac- teristic ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed and flanked at its base by a pair of strong pre- endopodal shields which abut the sternal shield. Sternal shield characteristically ornate (Text-fig. 125) and bearing three pairs of simple setae. Genital shield approximately in the form of a trapezium, heavily sclerotized and ornate, and bearing a single pair of simple setae. Ventri-anal shield broader than long (88-96 /i long X 172-180 /* wide), concave anteriorly, strongly ornate, and bearing nine simple setae. Between the genital and ventri-anal shields lies a pair of narrow sclerotized platelets. Stigma situated between coxae III and IV and peritreme not extending posterior to the stigma. Peritrematal shield well developed and extending posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 126) three-pronged, median prong broad and triangular, lateral prongs small, inconspicuous. Dentition of chelicera as in Text-fig. 127. Leg I (245-255 fJi long) with the tarsus (88-94 /*) twice the length of the tibia (43-45 fi). Setae on tarsus I very fine, ambulacrum short, claws small. Setae of remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae, and ambulacra with three acuminate lobes (Text-fig. 128). MALE. Dorsal shield (276-283 fi long x 180-185 /* wide) with ornamentation and chaetotaxy similar to the female. Tritosternum of the same form as that of the female. The chaetotaxy and sclerotization of the venter are shown in Text-fig. 129. The gnathosoma and tectum are similar to those of the female. The spermatophoral process is about the same length as the movable digit of the chelicera (Text-fig. 130). THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 77 FIGS. 124-130. Sejus aciculatus sp. nov. Fig., 124 dorsal shield of female. Fig. 125, venter of female. Fig. 126, tectum of female. Fig. 127, chelicera of female. Fig. 128, tarsus II of female. Fig. 129, venter of male. Fig. 130, chelicera of male. 78 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Leg I (281 11 long) with the tarsus (81 /*) twice the length of the tibia (38 /*). The chaetotaxy and ambulacra of the legs are as in the female. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.124), allotype male (1959.1.20.125), seven female and three male paratypes (1959.1.20.126-133) from damp leaf litter at St. Ann, Mt. Diablo, 1,500 ft. and 2,200 ft., 3^.1956, collected by P. F. Bellinger. A male and female paratype have been returned to the collector. Sejus antillanus sp. nov. FEMALE. Dorsal shield (333-443 /* long x 195-278 /* wide) slightly attenuated in its posterior half, and entirely covered by a reticulated pattern (Text-fig. 131). " Anterior dorsal shield " with twenty-one pairs of simple setae, verticals almost contiguous and directed anteriorly. " Posterior dorsal shield" with fifteen pairs of simple setae ; J5 are the shortest. The distribution and relative lengths of the setae, and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with a reticulated area anteriorly and bearing three pairs of simple setae (Text-fig. 132). Genital shield truncated posteriorly, reticulated, and bearing a single pair of setae. Ventri- anal shield broader than long (76-101 ju, long x 101-114 /* wide), reticulated, with nine simple setae of which one pair is long. Between the genital and ventri-anal shields lies a row of four small narrow sclerotized plates, and behind these a pair of smaller plates. Stigma situated between coxae III and IV ; no post-stigmal prolongation of the peritreme. Peritrematal shield extending posterior to coxae IV. Metapodal shields small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 133) three-pronged. The dentition of the chelicera is shown in Text-fig. 134. Leg I (384-521 /* long) with the tarsus (86-139 /*) sn ghtly longer than the tibia (78-114 fi). Setae on tarsus and tibia fine, those on remaining segments stouter ; claws on leg I small. Tarsi II-IV with a pair of lanceolate setae and ambulacra with acuminate lobes (Text-fig. 135). Leg IV (532-797 /*) considerably longer than the idiosoma, with the tarsus 203-316 /* long. MALE. Dorsal shield (236-303 ji long x 147-203 /* wide) having similar chaeto- taxy and ornamentation to the female. The chaetotaxy and sclerotization of the venter are shown in Text-fig. 136. The gnathosoma and tectum are similar to those of the female. The spermatophoral process is about one-and-a-half times the length of the movable digit (Text-fig. 137) . Leg I (238-432 pi long) with the tarsus (76-116 fi) longer than the tibia (63-88 /*). Leg IV (448-658 JLI long) with the tarsus 170-253 ju,. The chaetotaxy and ambulacra of the legs are as in the female. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.134) and four paratype females (1959 . i . 20 . 135-138) from damp leaf litter, Mt. Diablo, St. Ann, 2,200 ft., 7.vi.i956 ; the allotype male (1959.1.20.139) and five paratypes THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) 79 137 FIGS. 131-137. Sejus antillanus sp. nov. Fig. 131, dorsum of female. Fig. 132, venter of female. Fig. 133, tectum of female. Fig. 134, chelicera of female. Fig. 135, tarsus II of female. Fig. 136, venter of male. Fig. 137, chelicera of male. 8o THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 2?$) (1959.1.20.140-144) from damp leaf litter in wood, Dolphin Head, Hanover, 1,780 ft., 13. iv. 1956 ; three paratypes (i^, 2$$) in moss, St. Thomas, John Cove Mountains, 2,300 ft., 25 . v . 1956 ; and ten paratypes (ig, 9$$) (1959 . i . 20 . 145-154) from leaf litter, Portland Gap, 8.iii.i956. All were collected by P. F. Bellinger. A male and two female paratypes have been returned to the collector. FIGS. 138-142. Sejus bellingeri sp. nov. Fig. 138, dorsum of female. Fig. 139, tectum of female. Fig. 140, tarsus I of female. Fig. 141, venter of male. Fig. 142, chelicera of male. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 81 Sejus bellingeri sp. nov. FEMALE. Dorsal shield (467-507 fi long x 299-310 /i wide) slightly attenuated in its posterior half, and covered by a reticulated pattern (Text-fig. 138). " Anterior dorsal shield ' ' with twenty-one pairs of simple setae ; verticals closely situated and directed anteriorly. " Posterior dorsal shield " with fifteen pairs of simple setae of which J5 are the shortest. The distribution and relative lengths of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. The chaetotaxy and sclerotization of the venter are essentially the same as in Sejus nodosus (p. 85). The ventri-anal shield measures 119-139 /* long X 132-156 fju wide. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 139) three-pronged, and chelicerae typical of the genus. Leg I (c. 589 fi long) with the tarsus (c. 157 /*) longer than the tibia (c. 126 /*). Tarsus I with a strong erect seta dorsally (Text-fig. 140), remaining setae fine ; tarsal claws small but conspicuous. Tarsi II-III with a pair of lanceolate setae (tarsus IV has only one such seta), ambulacra with three acuminate lobes. Leg IV (c. 913 /*) longer than the idiosoma. Tarsus IV (303-341 /JL) markedly elongate and three times the length of the tibia. MALE. Dorsal shield (347 ju, long x 228 JJL wide) not attenuated in its posterior half, lightly reticulated. The chaetotaxy of the dorsal shield is the same as that of the female. The chaetotaxy and ornamentation of the venter are shown in Text-fig. 141. The gnathosoma and tectum are similar to those of the female. Spermatophoral process about twice the length of the movable digit of the chelicera (Text-fig. 142). Leg I (481 IJL long) with the tarsus (126 /*) longer than the tibia (99 fi). Tarsus I with a strong erect seta dorsally as in the female. Leg IV (675 /* long) longer than the idiosoma, and with the tarsus (265 /*) three times the length of the tibia (88 /*). LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.155), allotype male (1959.1.20.156) and one paratype female from pine and hardwood litter, St. Thomas, Portland Gap, 5,500 ft., 8.iii.i956, collected by P. F. Bellinger. The paratype has been returned to the collector. Sejus hulli sp. nov. FEMALE. Dorsal shield (359 /* long x 195 p wide) completely covered by a conspicuous reticulated pattern (Text-fig. 143). " Anterior dorsal shield " with twenty-one pairs of simple setae, verticals on a prominent protuberance and directed anteriorly. "Posterior dorsal shield" with fifteen pairs of simple setae, J5 are short. All setae arise from small protuberances. The distribution and relative lengths of the dorsal setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed with pilose laciniae. Sternal shield with a very faint 82 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) reticulated area medially, and bearing three pairs of simple setae (Text-fig. 144). Genital shield bearing a single pair of setae. Ventri-anal shield a little broader than long (94 /* long X 114 11 wide), reticulated and bearing nine simple setae. Between the genital and ventri-anal shields lie about six small sclerotized plates. Stigma situated between coxae III and IV ; no distinct post-stigmal prolongation of peri- treme ; peritrematal shield extending posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal seta on palptrochanter long, FIGS. 143-147. Sejus hulli sp. nov., female. Fig. 143, dorsum. Fig. 144, venter. Fig. 145, tectum. Fig. 146, chelicera. Fig. 147 tarsus II. whip-like. All setae on pedipalp simple. Tectum (Text-fig. 145) three-pronged. The dentition of the chelicera is shown in Text-fig. 146. Leg I (360 IJL) with the tarsus (101 /i) nearly twice the length of the tibia (56 /*). Setae on tarsus and tibia fine, those on remaining segments stouter. Claws on leg I minute, tarsus swollen in its distal third. Tarsi II-IV with a pair of lanceolate setae, and ambulacra with acuminate lobes (Text-fig. 147). MALE. Unknown. LOCALITY. Jamaica, B.W.I. A single female (1959.1.20.157) collected in Lycopodium at Fairy Glade, St. Andrew, 6.xii.i956, by P. F. Bellinger. THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 83 Sejus jamaicensis sp. nov. FEMALE. Dorsal shield (477 fi long x 310 jLt wide) concave posteriorly and covered with a light reticulated pattern (Text-fig. 148). " Anterior dorsal shield " with twenty-one pairs of simple setae ; verticals well separated, short, and arising from weak bases. " Posterior dorsal shield " with fifteen pairs of simple setae of which J5 are the shortest. The distribution of the dorsal setae and the ornamenta- tion of the dorsal shield are shown in the Text-fig. 148 FIGS. 148-152. Sejus jamaicensis sp. nov., female. Fig. 148, dorsum. Fig. 149, venter. Fig. 150 tectum. Fig. 151, chelicera. Fig. 152, tarsus II. Tritosternum well developed with pilose laciniae. Sternal shield with a little sculpturing and two circular areas anteriorly (Text-fig. 149), and bearing three pairs of simple setae. Genital shield with one pair of simple setae. Ventri-anal shield wider than long (114-119 p, long x 139-144 /* wide), bearing nine simple setae, and only lightly sculptured. Between the genital and ventri-anal shields lie a row of four narrow platelets and, posterior to these, two smaller ones. Stigma situated between coxae III and IV ; peritreme extending posterior to coxa IV. Metapodal plates very small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 150) three-pronged. The dentition of the chelicera is shown in Text-fig. 151. Leg I (924 fi long) with the tarsus (215 fi) only slightly shorter than the tibia (228 fi). Setae on tarsus, tibia and genu very fine ; those on remaining segments 84 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) stouter. Claws minute, sessile. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 152). MALE. Unknown. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.158) and one paratype female from wet leaves in spray zone, Cape Crow Falls, St. Andrew, 31 . iii . 1956, collected by P. F. Bellinger. The paratype has been returned to the collector. Sejus neborealis sp. nov. FEMALE. Dorsal shield (570 fi long X 330-340 p wide) entirely covered by a strong reticulated pattern, and with a sclerotized tubercle in the region of setae 154 FIGS. 153-157. Sejus neborealis sp. nov., female. Fig. 153, dorsum. Fig. 154, venter. Fig. 155, tectum. Fig. 156, chelicera. Fig. 157, tarsus II. J3-J4 (Text-fig. 153). "Anterior dorsal shield" with twenty-one pairs of simple setae, vertical setae contiguous and directed anteriorly. "Posterior dorsal shield " with fifteen pairs of simple setae, J4 situated on the margin of the sclerotized tubercle. The distribution and relative lengths of the dorsal setae, and the ornamentation of the dorsal shield are shown in the Text-fig. THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 85 Tritosternum well developed with pilose laciniae. Sternal shield with a small median reticulated area anteriorly, and bearing three pairs of short simple setae (Text-fig. 154). Genital shield slightly convex posteriorly and bearing one pair of simple setae. Ventri-anal shield broader than long (172-177 p long x 240-253 /* wide), wholly reticulated and bearing nine simple setae. Between the genital and ventri-anal shields lie four sclerotized platelets. Stigma situated between coxae III and IV with the posterior prolongation of the peritreme extending in line with the middle of coxa IV. Peritrematal shield well developed, extending posterior to coxa IV, and crenate externally. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 155) three-pronged. The dentition of the chelicera is shown in Text-fig. 156. Leg I (734-774 }i long) with the tarsus (228-243 /*) approximately twice the length of the tibia (114-126 /*). Some setae on tarsus I relatively long and standing at right angles to the segment ; ambulacrum short, claws small (and inconspicuous) ; setae of remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae and ambulacrum with three acuminate lobes (Text-fig. 157). MALE. Unknown. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.158) and three paratype females (1959.1.20.159-160) collected in moss on concrete, Coolsbade, Portland, 17 . vii . 1956, by P. F. Bellinger. One paratype has been returned to the collector. Sejus nodosus sp. nov. FEMALE. Dorsal shield (518-529 /* long x 310 /* wide) slightly attenuated in its posterior half, and covered by a reticulated pattern (Text-fig. 158). There is a pronounced sclerotized tubercle in the region of setae J3-J4- " Anterior dorsal shield " with twenty-one pairs of simple setae ; vertical setae closely situated and anteriorly directed. "Posterior dorsal shield" with twenty-one pairs of simple setae, of which J5 are the shortest. The distribution and relative lengths of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed and with pilose laciniae. Sternal shield plain and bearing three pairs of simple setae (Text-fig. 159). Genital shield slightly convex posteriorly, reticulated, and bearing a single pair of setae. Ventri-anal shield wider than long (114-116 fi long X 134-144 /* wide), lightly sculptured and bearing nine simple setae. Between the genital and ventri-anal shields lies a row of two to four narrow sclerotized platelets. Stigma situated between coxae III and IV ; peritreme not extending posterior to the stigma. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 160) three-pronged, and chelicerae typical of the genus. Leg I (c. 620 fi long) with the tarsus (165 /*) longer than the tibia (126 /*). Tarsus I with a strong erect spine dorsally, remaining setae fine ; tarsal claws small but 86 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) conspicuous. Tarsi II-III with a pair of acuminate lobes (Text-fig. 161.) Leg IV is about 913 fi long, with the tarsus 392 /* long, markedly elongate, and three times as long as the tibia. MALE. Unknown. 158 FIGS. 158-161. Sejus nodosus sp. nov., female. Fig. 158, dorsum. Fig. 159, venter. Fig. 160, tectum. Fig. 161, tarsus II. LOCALITY. Jamaica, B.W.I. The holotype female (1959.1.20.161) and one paratype from moss on concrete, Coolsbade, Portland, I7.vii.i955, collected by P. F. Bellinger. The paratype has been returned to the collector. Sejus browningi sp. nov. FEMALE. Dorsal shield (330-354 /* long x 190-210 /i wide) very heavily reticu- lated (Text-fig. 162). "Anterior dorsal shield" with twenty-one pairs of setae; " posterior dorsal shield" with fifteen pairs of setae. Dorsal setae J5 are short and simple, but the remainder are lanceolate (Text-fig. 163). The highest part of the dorsal shield is in the region of setae Ji-j3- The vertex is produced anteriorly. The distribution and relative lengths of the dorsal setae are shown in Text-fig. 162. THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) 87 Tritosternum with a pair of pilose laciniae. Sternal shield with a narrow longi- tudinal reticulated area anteriorly (Text-fig. 164) and bearing three pairs of simple setae. Genital shield broad, truncated posteriorly, lightly striated and bearing a single pair of simple setae. Ventri-anal shield wider than long (99-101 /* long X 121-139 fi wide), lightly reticulated anteriorly, and bearing nine simple setae. A row of five narrow sclerotized plates lies on the membrane between the genital and ventri-anal shields. Stigma situated between coxae III and IV, with the peritreme 164 FIGS. 162-167. Sejus browningi sp. nov., female. Fig. 162, dorsal shield. Fig. 163, dorsal seta. Fig. 164, venter. Fig. 165, tectum. Fig. 166, chelicera. Fig. 167, tarsus II. extending only a short distance posterior to the stigma. Peritrematal shield extending posterior to coxa IV. Metapodal shields triangular and situated close to the peritrematal shield. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 165) three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 166. Leg I (320-340 fi long) with the tarsus (88 fi) longer than the tibia (58 p). Setae on tarsus I very fine ; those on remaining segments somewhat stouter. Tarsus I is swollen in its distal third ; without ambulacrum. Tarsi II-IV with a pair of lanceolate setae, and ambulacra with lateral lobes acuminate ; median lobe strongly produced and rounded apically (Text-fig. 167). 88 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) MALE. Unknown. LOCALITY. Sikkim and Argentina. The holotype female (1959.1.20.162) and two paratype females (1959.1.20.163-164) from leaves and leaf -mould in a wood on steep hillside, Chungtang, Sikkim, 5,120 ft., i6.ii.i952, collected by Dr. Theresa Clay ; and another paratype female (1959.1.20.165) from rotting vegetable matter, Tucuman City, Argentina, 500 m., January, 1953, collected by Dr. P. Wygodzynsky. Sejus ornatus sp. nov. FEMALE. Dorsal shield (395-400 fi long x 245-253 fi wide) with lateral incisions between setae 82 and 83, and covered by a network of small tubercles (Text-fig. 1 68). Highest part of dorsal shield in the region of setae Ji-j3- " Anterior dorsal shield " with twenty- two pairs of simple setae ; " posterior dorsal shield " with fifteen pairs of simple setae. Setae J5 are long. Vertex of dorsal shield produced ante- 169 FIGS. 168-173. Sejus ornatus sp. nov., female. Fig. 168, dorsum. Fig. 169, venter. Fig. 170, vertex. Fig. 171, tectum. Fig. 172, chelicera. Fig. 173, tarsus II. THE PLATYSEIINAE ' (MESOSTIGMATA : ACEOSEJIDAE) 89 riorly. The distribution and relative lengths of the dorsal setae and the ornamenta- tion of the dorsal shield are shown in the Text-fig. Tritosternum well developed, with pilose laciniae. Sternal shield with a circular impression antero-medially and bearing three pairs of simple setae (Text-fig. 169). Genital shield broad, truncated posteriorly, and bearing a single pair of setae. Ventri-anal shield wider than long (118-139 fi long x 164-177 fi wide), ornamented with small tubercles and punctations and bearing nine simple setae. Stigma situated between coxae III and IV ; peritremes not joined anteriorly (Text-fig. 170), and not extending posterior to the stigmata. Outer margin of peritreme crenate ; peritrematal shield extending to the posterior level of coxa IV. Metapodal shields narrow and transversely situated. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 171) three-pronged, typical of the genus. The dentition of the chelicera is shown in Text-fig. 172. Leg I (360-366 fji long) with the tarsus (103-113 /*) longer than the tibia (58-63 /*). Setae on tarsus I very fine ; those on remaining segments stouter. Tarsi II-IV with a pair of lanceolate setae, and ambulacra with lateral lobes acuminate, median lobe rudimentary (Text-fig. 173). Tarsus I slightly swollen at its distal half, pulvillus and claws small. MALE. Unknown. LOCALITY. Argentina. The holotype female (1959.1.20.166) and three para- types (1959.1.20.167-169) from horse dung and decomposing vegetable matter in forest, Quebrada de los Sosa, Tafi, Tucuman, 1,300 m., 17. v. 1953, collected by Dr. P. Wygodzynsky. Sejus tuberculatus sp. nov. FEMALE. Dorsal shield (477-518 fi long x 310-330 /JL wide) minutely punctured and slightly concave posteriorly (Text-fig. 174). "Anterior dorsal shield" with twenty-one pairs of simple setae. " Posterior dorsal shield " with fifteen pairs of simple setae ; J5 are the shortest. The distribution and relative lengths of the setae, and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum well developed and with pilose laciniae. Sternal shield with faint sculpturing, and bearing three pairs of simple setae (Text-fig. 175). Genital shield flask-shaped, and bearing a single pair of setae. Ventri-anal shield almost circular in outline (130-144 fi long x 144-147 /i wide) and bearing nine simple setae. Between the genital and ventri-anal shields lies a transverse row of about three narrow plates, and posterior to these lie two small plates. Stigma situated between coxae III and IV ; peritreme and peritrematal shields extending posterior to coxa IV. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 176) three-pronged. The dentition of the chelicera is shown in Text-fig. 177. Leg I (794-804 fi long) with the tarsus (162-177 (JL) shorter than the tibia (182- ZOOL. 6, NO. 2. 7 90 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 197 /*). Setae on tarsus, tibia and genu fine ; those on other segments stouter ; tarsal claws minute, inconspicuous. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with three acuminate lobes (Text-fig. 178). MALE. Unknown. I %t -^ i\t I T>, A FIGS. 174-178. Sejus tuberculatus sp. nov., female. Fig. 174, dorsum. Fig. 175, venter. Fig. 176, tectum. Fig. 177, chelicera. Fig. 178, tarsus II. LOCALITY. Argentina. The holotype female (1959 . 1 . 20 . 170) and two paratype females (1959.1.20.171.172) from decaying vegetable matter in forest, Quebrada de los Sosa, Tafi, Tucuman, 13. v. 1953, 1,300 m., collected by Dr. P. Wygodzynsky. The following species also appear to belong to the genus Sejus, but with the exception of Episeius groenlandicus, have not been examined by us : Acarus tendens Schrank, 1803. Fauna Boica, 3 : 209. Oudemans (1929) con- siders this species to be synonymous with Platyseius subglaber (Ouds.). Gamasus scabriculus Nordmann, 1832. Mikr. Beitr. Naturg. wirbell. Thiere : 85. Epicrius glaber Berlese, 1886. A. M.S., 30, No. 9. Italy. Hypoaspis scutulis Banks, 1914. Psyche, 21 : 161. Brazil. Lasioseius borealis var. temperatus Berlese, 1916. Redia, 12 : 34. Palermo. Lasioseius grandis Berlese, 1916. Tom. cit. : 34. Piemonte, Venice. Lasioseius parapodicus Berlese, 1916. Tom. cit. : 35. Java. Lasioseius similis Berlese, 1916. Tom. cit. : 35. Italy. This species may be a THE PLATYSEIINAE (MESOSTIGMAT A : ACEOSEJIDAE) 91 synonym of Sejus serratus (Halbert). Lasioseius mutilus Berlese, 1916. Tom. cit. : 37. Florence. Lasioseius (Cheiroseius) alpestris Berlese, 1916. Tom. cit. : 41. Italy. Episeius major incisus Willmann, 1938. Ann. Hist. not. Mus. Hung., 31 : 168. Hungary. Episeius groenlandicus Haarlev, 1942. Medd. Gronland, 128 : 18. Greenland. Episeius aequaLis Schweizer, 1949. Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 64. Switzerland. Episeius handschini Schweizer, 1949. Tom. cit. : 71. Switzerland. Episeius salicorniae Willmann, 1949. Veroff. Mus. Nat. Bremen, No. lA : 118. Poland. Episeius longipes Willmann, 1951. SB. ost. Akad. Wiss., Abt. I, 160 : 112. Austria. Platyseius mackerassae Womersley, 1956. /. Linn. Soc. Lond. (Zool.) 42 : 552. Queensland. Genus ZERCONOPSIS Hull Zerconopsis Hull, J. E., 1918, Trans, nat. Hist. Soc. Northumb., 5 : 65. Platyseiine mites with the "posterior dorsal shield" bearing fourteen pairs of setae comprising five pairs in the J and the Z series but only four pairs in the S series (82 being situated on the lateral interscutal membrane). Three to five pairs of dorsal setae stout and paddle-like (Text-figs. 179, 192). Lateral margins of the shield entire or incised anterior to Si. Sternal shield in the female with three pairs of setae ; metasternals situated on platelets. Genital shield wedge shaped, genital setae situated on or off the shield. Ventri-anal shield large and bearing from three to seven pairs of setae in addition to the three setae normally associated with the anus. Male with sterniti-genital and free ventri-anal shield. Peritreme without post-stigmatic process ; peritrematal shield poorly developed posterior to coxa IV. Chaetotaxy of the venter of the gnathosoma and of the pedipalps typical of the subfamily. Chelicerae chelate-dentate, movable digit in the male with a short spermatophoral process ; tectum basically three-pronged. Legs I-IV with a well- developed ambulacrum ; median lobes of pulvilli rounded apically. Type species Gamasus remiger Kramer, 1876 KEY TO SPECIES ADULTS 1. Dorsal shield with distinct lateral incisions anterior to Si (Text-figs. 184 and 197) 2 Lateral margins of the dorsal shield entire ....... 3 2. Genital shield in the female narrow, without setae (Text-fig. 198). Labrador Z. I abradorensis sp. n. (p. 99) Genital shield in the female broad, with a pair of setae (Text-fig. 185). Switzerland Z. muestairi (Schweizer) (p. 94) 3. Dorsal shield with five pairs of paddle-like setae (Text-fig. 192) ; genital shield without setae (Text-fig. 193). Hungary . . Z. decemremiger sp. n. (p. 96) 92 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Dorsal shield with three pairs of paddle-like setae (Text-figs. 179 and 189) ; genital shield with or without setae .......... 4 4. Genital shield with a pair of setae (Text-fig. 180) ; with one pair of large, subtriangular metapodal shields. Europe ..... Z. remiger (Kramer) (p. 92) Genital shield narrow, without setae (Text-fig. 190) ; with two pairs of metapodal shields. England ....... Z. michaeli sp. n. (p. 95) Zerconopsis remiger (Kramer) Gamasus regimer Kramer, P., 1876, Arch. Naturg., Ameroseius bispinosus Berlese, A., 1910, Redia, 6 : 42 : 93. 6 : 253. FEMALE. Dorsal shield (570-630 /* long X 310-382 /* wide) covered antero- laterally with a distinct reticulated pattern ; posterior and central areas with a lighter network (Text-fig. 179). " Anterior dorsal shield " with twenty- two pairs of setae of which one pair is paddle-like, the remainder simple ; verticals short, each on a tubercle. " Posterior dorsal shield " with fourteen pairs of setae of which two pairs are paddle-like and the remainder simple ; J5 are the shortest. The distribu- tion and relative lengths of the dorsal setae and the ornamentation of the dorsal shield are shown in Text-fig. 179. Tritosternum with a narrow base and pilose laciniae, and flanked by a narrow pair of pre-endopodal shields. Sternal shield with a fine granular pattern and bearing three pairs of simple setae (Text-fig. 180). Genital shield also with granular markings, and bearing one pair of setae. Ventri-anal shield wider than long (220- 223 IJL long x 225-303 [A wide), heavily reticulated and bearing thirteen simple setae. Between the genital and ventri-anal shields lie six platelets. Stigma situated between coxae III and IV with the peritreme not extending posterior to the stigma. Metapodal plates conspicuous. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of the genus. The dentition of the chelicera is shown in Text-fig. 181. Leg I (379-400 fi long) with the tarsus (99-103 /*) longer than the tibia (63-68 /*) ; claws small, pulvillus short. Tarsus I smooth with fine setae ; tibia, genu and femur with majority of setae arising from strong tubercles. Tarsi II-IV with a pair of lanceolate setae ; ambulacra with median lobes of pulvilli rounded apically. MALE. Dorsal shield (530-550 /* long x 299-310 JJL wide) with chaetotaxy and ornamentation similar to the female. Sterniti-genital shield with five pairs of simple setae. Ventri-anal shield (Text-fig. 182) reticulated anteriorly and bearing seventeen simple setae. The gnathosoma, tectum, pedipalps and legs are similar to those of the female. Leg I (c. 420 fi long) with the tarsus (99 fi) and the tibia (68 /*). Sperma- tophoral process about a third as long again as the movable digit of the chelicera (Text-fig. 183). DISTRIBUTION. Kramer (1876) described this species from under fallen leaves without giving a locality, and Berlese (1910) described Ameroseius bispinosus from rotting leaves at Palermo. We have examined specimens from Nadap, Herkules- THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) 93 fiirdo, and Gyertyanliget in Hungary, and British specimens from Goyt Valley, Cheshire, 24. xi. 1940, collected by H. Britten; Rydal Water, Westmorland, 29. xi. 1954, collected by M. E. Bacchus ; Harefield, Middlesex, 22^.1956, collected by A. H. G. Alston; and Avonmouth, Gloucestershire, n.v.i957, collected by P. N. Lawrence. 183 FIGS. 179-183. Zerconopsis remiger (Kramer). Fig. 179, dorsum of female. Fig. 180, venter of female. Fig. 181, chelicera of female. Fig. 182, ventri-anal shield of male. Fig. 183, chelicera of male. 94 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) Zerconopsis muestairi (Schweizer) comb. nov. Lasioseius mustairi Schweizer, J., 1949, Rec. Rech. sci. Pare Nat. Suisse N.F. 2 : 50. FEMALE. Dorsal shield (585-630 fi long X 315-378 /* wide) with a lateral incision anterior to seta Si ; lightly reticulated anteriorly and entirely granular (Text-fig. 184). " Anterior dorsal shield " with nineteen pairs of setae of which one pair is paddle-like, remainder simple ; verticals relatively long and well separated. " Pos- terior dorsal shield " with fourteen pairs of setae of which two pairs are paddle-like, FIGS. 184-188. Zerconopsis muestairi (Schweizer). Fig. 184, dorsum of female. Fig. 185, venter of female. Fig. 186, tectum of female. Fig. 187. ventri-anal shield of male. ] { 188, chelicera of male. THE PLATYSEIINAE (MESOSTIGMATA : ACEOSEJIDAE) 95 the remainder simple ; J5 are the shortest. The distribution and relative lengths of the dorsal setae and the form of the dorsal shield are shown in the Text-fig. Tritosternum with a narrow base and pilose laciniae. Sternal shield with faint sculpturing and bearing three pairs of simple setae (Text-fig. 185). Genital shield also with faint markings, and bearing a single pair of setae. Ventri-anal shield wider than long (180-223 /* l n g X 240-270 /* wide), reticulated in its anterior half and bearing thirteen simple setae. Between the genital and ventri-anal shields lie six platelets. Stigma situated between coxae III and IV with the peritreme not extending posterior to the stigma. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 186) three-pronged and typical of the genus. The chelicerae are figured by Schweizer (1949). Leg I (432-504 IJL long) with the tarsus (118-132 /*) longer than the tibia (68-75 /*). Claws on leg I distinct, pulvillus small ; tarsal setae fine, those on other segments stouter. Tarsi II-IV with a pair of lanceolate setae ; median lobes of pulvilli rounded apically. MALE. Dorsal shield (467-498 /* long x 252-289 ju, wide) with chaetotaxy and ornamentation similar to the female. Sterniti-genital shield with five pairs of simple setae. Ventri-anal shield with nineteen simple setae (Text-fig. 187). The gnathosoma, tectum, pedipalps and legs are typical. Leg I (326-332 fi long) with the tarsus 108-116 /* and the tibia 63-65 /*. The spermatophoral process and chelicera are shown in Text-fig. 188. LOCALITY. This species is known only from the Swiss National Park. We have examined five females and three males from the type series. Zerconopsis michaeli sp. nov. FEMALE. Dorsal shield (498 fi long x 253 fi wide) with a heavy distinct pattern of depressions ; lateral margins irregular, posterior margin crenate (Text-fig. 189). " Anterior dorsal shield " with nineteen pairs of setae of which one pair is paddle- like, the remainder simple ; verticals on tubercles and directed anteriorly. " Pos- terior dorsal shield " with fourteen pairs of setae of which two pairs are paddle-like, the remainder simple. The distribution of the setae and the structure and orna- mentation of the dorsal shield are shown in the Text-fig. Tritosternum with a narrow base and pilose laciniae. Sternal shield plain, bearing three pairs of simple setae (Text-fig. 190). Genital shield narrow, setae off the shield. Ventri-anal shield wider than long (160 /* long X 212 fi wide), with a pattern of reticulations and dots, and bearing eleven setae of which the post-anal is long and stout. There are at least two narrow sclerotized platelets between the genital and ventri-anal shields. Stigma situated between coxae III and IV ; no post-stigmal prolongation of the peritreme. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 191) three-pronged and typical of the genus. Chelicerae typical. 9 6 THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) Leg I (387 fi long) with the tarsus (101 /*) almost twice the length of the tibia (58 X>. Legs II-IV incomplete in the unique specimen, but some of the setae arise from tubercles. MALE. Unknown. LOCALITY. A single female (1930.8.25.1840) from "England" in the A. D. Michael Collection and labelled " Sejus remiger Kramer ". 190 FIGS. 189-191. Zerconopsis michaeli sp. nov., female. Fig. 189, dorsal shield. Fig. 190, venter. Fig. 191, tectum. Zerconopsis decemremiger sp. nov. FEMALE. Dorsal shield (590-600 /* long x 360-382 /* wide) with a dense pattern of depressions, largest around the anterior margin and becoming smaller posteriorly (Text-fig. 192). " Anterior dorsal shield " with twenty pairs of setae of which two pairs are paddle-like, the remainder simple ; verticals separated by about three times the diameter of their bases. " Posterior dorsal shield " with fourteen pairs of setae of which three pairs are paddle-like ; J5 are the shortest. The distribution of the setae and the ornamentation of the dorsal shield are shown in the Text-fig. Tritosternum with a narrow base and pilose laciniae. Sternal shield with faint THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 97 I 14 FIGS. 192-196. Zerconopsis decemremiger sp. nov. Fig. 192, dorsal shield of female. Fig. 193, venter of female. Fig. 194, tectum of female. Fig. 195, ventri-anal shield of male. Fig. 196, chelicera of male. 98 THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) reticulations and bearing three pairs of simple setae (Text-fig. 193). Genital shield reticulated ; genital setae off the shield. Ventri-anal shield wider than long (205- 220 IJL long X 253-263 fi wide), with a conspicuous pattern of small depressions and bearing eleven simple setae. Between the genital and ventri-anal shields lie four small plates. Stigma situated between coxae III and IV ; peritreme without post-stigmal prolongation. Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum three-pronged and typical of Zerconop- sis (Text-fig. 194) : chelicerae also typical. Leg I (434 /* long) with the tarsus (m-ii6 /*) longer than the tibia (68-70 /*). Claws on leg I small ; tarsal setae fine, those on remaining segments somewhat stouter and arising mainly from small tubercles. Tarsi II-IV with a pair of lanceo- late setae. Ambulacra with median lobes of pulvilli rounded apically. MALE. With dorsal shield (507-550 /* long x 289-330 fi wide) essentially the same as the female. Sterniti-genital shield with five pairs of simple setae. Ventri- anal shield (Text-fig. 195) similar in ornamentation to the female and bearing nineteen simple setae. The gnathosoma, tectum, pedipalps and legs are similar to those of the female. Leg I (434 fi long) with the tarsus (111-113 A) longer than the tibia (68-70 /*). The spermatophoral process and chelicera are shown in Text-fig. 196. LOCALITY. Hungary. The holotype female (1959.1.20.177), allotype male (1959.1.20.178), four female and two male paratypes (1959.1.20.179-184) from Herkulesfurdo, April- June, 1938 ; and four female and two male paratypes from Gyertyanliget, 6-i7th August, 1940, sent to us by Dr. J. Balogh, Budapest. Zerconopsis labradorensis sp. nov. FEMALE. Dorsal shield (581-609 fi long x 299-320 /* wide) with a lateral incision anterior to seta Si, reticulated antero-laterally and the remainder granulate (Text- fig. 197). " Anterior dorsal shield " with twenty-one pairs of setae of which one pair is paddle-like, the remainder simple ; vertical setae with their bases two diameters apart. " Posterior dorsal shield " with fourteen pairs of setae of which two pairs are paddle-like ; J5 are the shortest. The distribution of the setae and the form of the dorsal shield are shown in the Text-fig. Tritosternum with a narrow base and a pair of pilose laciniae. Sternal shield almost without markings, and bearing three pairs of simple setae (Text-fig. 198). Genital shield plain ; setae off the shield. Ventri-anal shield wider than long (177- 215 /* long X 253-260 fi wide), reticulated anteriorly and granulate posteriorly, and bearing fifteen simple setae. Between the genital and ventri-anal shields lie six small sclerotized plates. Stigma between coxae III and IV ; peritreme without post-stigmal prolongation (Text-fig. 199). Metapodal plates small. Venter of gnathosoma with rostral and internal palptrochanter setae long, whip- like. All setae on pedipalp simple. Tectum (Text-fig. 200) three-pronged ; cheli- cerae typical (Text-fig. 201). THE PLATYSEIINAE (MESOSTIGM AT A : ACEOSEJIDAE) 99 FIGS. 197-204. Zerconopsis labradorensis sp. nov. Fig. 197, dorsum of female. Fig. 198, venter of female. Fig. 199, peritrematal shield of female. Fig. 200, tectum of female. Fig. 201, chelicera of female. Fig. 202, tarsus II of female. Fig. 203, venter of male. Fig. 204, chelicera of male. loo THE PLATYSEIINAE (MESOSTIGM ATA : ACEOSEJIDAE) Leg I (462-502 fjb long) with the tarsus (124-134 fi) twice the length of the tibia (65-73 fi) ; tarsal claws small ; setae fine, those on remaining segments shorter and stouter. Tarsi II-IV (Text-fig. 202) with a pair of lanceolate setae ; ambulacra with median lobes of pulvilli rounded apically. MALE. Dorsal shield (488 /* long x 238 fi wide) essentially the same as in the female. Sterniti-genital shield with five pairs of simple setae. Ventri-anal shield (Text-fig 203) similar in ornamentation to the female and bearing nineteen simple setae. The gnathosoma, tectum, pedipalps and legs are similar to those of the female. Leg I (462 ju, long) with the tarsus (124 fi) twice as long as the tibia (65 /*). The spermatophoral process of the chelicera is shown in Text-fig. 204. LOCALITY. Labrador. The holotype female (1959.1.20.185), allotype male (1959.1.20.186) and six paratype females (1959.1.20.187-192) from wet willow humus, West Carter Basin, 4-17 . viii . 1958, collected by P. N. Lawrence during the British Schools' Exploring Society Expedition. SUMMARY 1. This paper deals with the external morphology and classification of the Platy- seiinae. Four genera, namely, Sejus C. L. Koch (syn. Cheiroseius, Episeius and Episeiella), Platyseius Berlese, Zerconopsis Hull, and Plesiosejus gen. nov. are recognized. Keys are given to the species represented in the Collections of the British Museum (Natural History). 2. The following new synonymy is introduced : Lasioseius (Platyseius) capillatus Berl. = Hypoaspis subglabm Oudms. Paraseius tenuipes Halbert = Ameroseius italicus Berl. Episeiella heteropoda Willm. = Sejus viduus C. L. Koch Episeius montanus Willm. = Ameroseius borealis Berl. Episeius ovaspini Schweizer = Lasioseius (Episeius} glaber var. curtipes Halbt. Parasejus glaber var. minor Tragardh = Epicrius laelaptoides Berl. Lasioseius (Episeius} sphagni Halbt. = Epicrius laelaptoides Berl. 3. The following twenty-seven new species are described and figured : Platyseius jamaicensis ; Platyseius spinosus ; Plesiosejus horridus ; Sejus aciculatus ; S. alstoni ; S. antillanus ; S. areolatus ; S. bellingeri ; S. browningi ; S. cassiteridum ; S. clayi ; S. dromadis ; S. hulli ; S. jamaicensis ; S. kennedyi ; S. neborealis ; S. nepalensis ; S. nodosus ; S. ornatus ; 5. parbatensis ; S. phalangioides ; 5. signatus ; S. spinipes ; 5. tuberculatus ; Zerconopsis decemremiger ; Z. labradorensis and Z. michaeli. REFERENCES BERLESE, A. 1882-1903. Acari, myriapoda et scorpiones hucusque in Italia reperta. Portici et Padua. 1904. Acari nuovi II. Redia, 1 : 258-280. 1905. Acari nuovi V. Ibid. 2 : 231-238. 1910. Acari nuovi VI. Ibid. 6 : tav. 19, figs. 35, 350. 1910. Lista di nuove specie e nuovi generi di Acari. Ibid. 6 : 243-271. THE PLATYSEIINAE (MESOSTI GM AT A : ACEOSEJIDAE) 101 BERLESE, A. 1913. Acari nuovi VII-VIII. Ibid. 9 : 77-111. 1916. Centuria prima di Acari nuovi. Ibid. 12 : 19-67. BUITENDIJK, A. M. 1945. Voorloopige Catalogus van de Acari in de Collectie-Oudemans. Zool. Med. 23 : 281-391. EVANS, G. O. 1957. An introduction to the British Mesostigmata (Acarina) with keys to families and genera. /. Linn. Soc. Land. (Zool.) 43 : 203-259. 1958. A revision of the British Aceosejinae (Acarina : Mesostigmata). Proc. zool. Soc. Lond. 131 : 177-229. FRANZ, H. 1943. Die Landtierwelt der Mittleren Hohen Tauern. Denkschr. Akad. Wiss. Wien, 107 : 79-119. HALBERT, J. N. 1915. Clare Island Survey Part 3gii Acarinida ii Terrestrial and marine Acarina. Proc. R. Irish Acad. 31 : 45-136. 1923. Notes on Acari with descriptions of new species. /. Linn. Soc. Lond. (Zool.) 35 : 363-392. HIRSCHMANN, W. 1957- Acarologie Gangsystematik der Parasitiformes Teil i. Rumpf- behaarung und Ruckenflachen. Schriftenreihe fur vergleichende Milbenkunde. Fiirth/ Bay : 1-20 i-v, pis. 1-26. HULL, J. E. 1918. Terrestrial acari of the Tyne Province. Trans, nat. Hist. Soc. Northumb. (n.s.) 5 : 13-88. KOCH, C. L. 1836-1844. Deutschlands Crustacean, Myriapoden und Arachniden. Regens- burg. 1837-1850. Ubersicht des Arachnidensy stems. Niirnberg. KRAMER, P. 1876. Zur Naturgeschichte einiger Gattungen aus der Familie der Gamasiden. Arch. Naturg. 42 : 46-105. OUDEMANS, A. C. 1902. Acariden von Borkum und Wangeroog. Abh. nat. Ver. Bremen, 18 : 77-98. 1903. Acarologische Aanteekeningen, VI. Ent. Ber. 1 : 83-88. 1929. Kritisch Historisch Overzicht der Acarologie, Leiden, 2 : 87 1936. Ibid 3 A : 1-430. SCHWEIZER, J. 1922. Beitrag zur Kenntnis der terrestrischen Milbenfauna der Schweiz. Verh. Naturf. Ges. Basel, 33 : 23-112. 1924. Beitrage zur Kenntnis der Tierwelt norddeutscher Quellgebiete. Acarina (land- milben). Arch, fur Hydrobiol. 15 : 125-132. 1949. Die Landmilben des Schweizerischen Nationalparkes. i. Parasitiformes. Res. Rech. sci. Pare Nat. Suisse, N.F. 2 : 1-99. SELLNICK, M. 1944. Zercon C. L. Koch. A cari-B latter fur Milbenkunde. Konigsberg, No. 5 : 30-41. TRAGARDH, I. 1910. Acariden aus dem Sarekgebirge. Naturw. Untersuch. Sarekgeb. 4 : 375- 586. 1946. Outlines of a new classification of the Mesostigmata (Acarina) based on comparative morphological data. Acta Univ. Lund, N.F. Avd. 2, 42, 4 : 1-32. VITZTHUM, H. G. 1931. Terristrische Acarinen (unter Ausschluss der Oribatiden und Ixodiden) der Deutschen Limnologischen Sunda-Expedition. Arch, fur Hydrobiol., Suppl., 9 : 59- 134- WILLMANN, C. 1938. Beitrag zur Kenntnis der Acarofauna des Komitates Bars. Ann. Hist. nat. Mus. Hung. 31 : 144-172. 1949. Beitrage zur Kenntnis des Salzgebietes von Ciechocinek. Veroff. Mus. Nat. Bremen, No. lA : 106-135. 1952. Die Milbenfauna der Nordsee-Insel Wangerooge. Veroff. Inst. Meeresf. Bremer- haven, 1 : 139-186. 1953. Neue Milben aus den ostlichen Alpen. SB. ost. Akad. Wiss. 162 : 449-5 J 9- 9 6t BRITISH SLUGS (PULMONATA ; TESTACELLIDAE, ARIONIDAE, LIMACIDAE) H. E. QUICK BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 3 LONDON: 1960 BRITISH SLUGS (PULMONATA ; TESTACELLIDAE, ARIONIDAE, LIMACIDAE) BY H. E. QUICK Pp. 103-226 ; 2 Plates ; 19 Text-figures ; 23 Maps BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 3 LONDON: 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 3 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March, 1960 Price Forty- five Shillings BRITISH SLUGS (PULMONATA ; TESTACELLIDAE, ARIONIDAE, LIMACIDAE) By H. E. QUICK CONTENTS INTRODUCTION ......... Arrangement ........ Acknowledgments ........ NATIVE AND NATURALIZED SPECIES . Order Stylommatophora . ...... Superfamily Oleacinacea ..... Family Testacellidae ..... TESTACELLA Draparnaud Testacella maugei Ferussac Testacella haliotidea Draparnaud Testacella scutitlum Sowerby Superfamily Endodontacea ..... Family Arionidae ...... Subfamily Arioninae ..... GEOMA LA C US Allman Geomalacus maculosus Allman ARION Ferussac .... Arion intermedius Normand . A rion fasciatus (Nilsson) Arion hortensis Ferussac Arion subfuscus (Draparnaud) Arion lusitanicus Mabille Avion ater ater (Linnaeus) Arion ater rufus (Linnaeus) . Superfamily Zonitacea ...... Family Limacidae ...... Subfamily Parmacellinae .... MILAX Gray Milax gagates (Draparnaud) . Milax cf. insularis (Lessona & Pollonera) Milax sowerbyi (Ferussac) Milax budapestensis (Hazay) ZOOL. 6, 3. 106 BRITISH SLUGS Page Subfamily Limacinae . . . . . .163 AGRIOLIMAX Morch 163 Agriolimax reticulatus (Miiller) . . . .164 Agriolimax agrestis (Linnaeus) . . . .170 Agriolimax laevis (Miiller) . . . . .172 Agriolimax caruanae Pollonera . . . .175 LIMAX Linnaeus '> .180 Limax tenellus Miiller . . . . . .180 Limax flavus Linnaeus . . . . .184 Limax cinereoniger Wolf . . . . .187 Limax maximus Linnaeus . . . . igi LEHMANNIA Heynemann . . . .193 Lehmannia marginata (Miiller) . . . ..194 GREENHOUSE ALIENS ......... 197 Lehmannia poirieri (Mabille) ........ 197 Limax nyctelius Bourguignat . . . . . . . . 200 FOSSIL SPECIES ........... 202 Limax modioliformis Sandberger ....... 202 ANATOMICAL TERMS .......... 203 DISTRIBUTION ........... 205 REFERENCES ........... 217 INDEX ............. 224 INTRODUCTION No comprehensive work on British slugs has appeared since John William Taylor's classic monograph (1902-07). Besides full information on their nomenclature, systematics and biology, this contains excellent coloured plates, depicting numerous varieties of the eighteen species of slugs which he recognized as British. Besides much information on these species which has become available since 1907, five more species, a subspecies, and two aliens established in British greenhouses must be added to the list. Taylor omitted Arion lusitanicus (p. 135), although it had already been recorded from Britain, and it has since reappeared. Milax budapestensis (p. 160), another species of Milax provisionally referred to M. insularis (p. 156) and Agriolimax caruanae (p. 175) have been recorded ; Agriolimax reticulatus (p. 164) has been shown to be specifically distinct from A. agrestis, and Arion ater rufus (p. 145) to be subspecifically distinct from A. ater. The alien species Lehmannia poirieri (p. 197) and Limax nyctelius (p. 202) have been found in Britain, but only in greenhouses. Two names used by Taylor must be changed : Arion circumscriptus to A.fasciatus (p. 127) and Limax arborum to Lehmannia marginata (p. 194). Two more recent works also deal with all British slugs, though less completely. Ellis (1926) gives monochrome figures but only occasional references to internal anatomy ; while Quick (1949) also gives monochrome figures, and information relating mainly to identification in the field. Among more general works, Kunkel (1916) on the biology of pulmonates, 0kland (1923) on the Arionidae of Norway, and Boycott (1934) on the habits of British land molluscs may be especially mentioned. Much information on courtship and copulation is given by Gerhardt (1933, 1935, 1939 and 1940), on the seasonal abundance and growth of slugs in gardens and fields BRITISH SLUGS 107 by Barnes & Weill (1945, with good coloured plates of garden species), and on their ecology by Thomas (1944) . For full synonymies, reference should be made to Kennard & Woodward (1926), and also to Taylor (1902-07) and Hesse (1926). Only the more important synonyms are quoted below. The group of naked pulmonates termed slugs is highly polyphyletic in origin. British slugs belong to the three families Testacellidae, Arionidae and Limacidae, which have evidently been derived from shelled Oleacinids, Endodontids and Zonitids respectively. Arrangement Information is given in the same sequence for each species (except that the shell is described before the external variations for Testacella) under the following headings, any of which may be omitted. Notes Radula First British record Alimentary system Diagnostic features Reproductive system External appearance Spermatophore External variation Mating Shell Development Pallial organs Behaviour Pedal gland Ecology Retractor muscles Distribution Nervous system Fossil record Jaw Material examined The anatomical terms used in the descriptions are explained by means of a glossary and semi-diagrammatic figures on p. 203. Distributions in the British Isles are shown on vice-comital maps (p. 205), as well as being briefly described under " distribution ". Acknowledgments I wish to thank the following. The Trustees of the British Museum for financial assistance, and Dr. H. W. Parker, C.B.E. for access to the zoological collections of the British Museum (Natural History). Mr. A. E. Ellis for preparing the maps. M. J. M. Gaillard for information and specimens from the Museum National d'His- toire Naturelle, Paris. Mr. I. C. J. Galbraith for editorial assistance. Dr. H. Loh- mander for information on Scandinavian slugs. Dr. O. E. Paget for information from the Naturliistorisches Museum, Vienna. Dr. W. J. Rees for valuable advice. Mr. H. O. Ricketts for much help in connection with material in the B.M. (N.H.), and for his care in preparing the bibliography and index. The late Mr. Hugh Watson for much helpful correspondence. The many correspondents who have helped from time to time by providing specimens or information, and especially Dr. W. Adam, Dr. C. O. van Regteren Altena, Mr. D. Aten, Dr. H. F. Barnes, Dr. A. J. Cain, Mr. A. E. Ellis, Dr. L. Forcart, Dr. W. O. Gregg, Dr. P. Makings, Dr. A. D. J. Meeuse, Dr. Nils Odhner, Dr. A. W. Stelfox, Mr. Abu Teira, Mrs. M. Vincent, and Dr. M. H. Williamson. io8 BRITISH SLUGS NATIVE AND NATURALIZED SPECIES Order STYLOMMATOPHORA Pulmonate mollusca bearing eyes at the apex of retractile tentacles. Superfamily OLEACINACEA Jaw rudimentary or absent, radular teeth aculeate, shell turreted or auriform. Carnivorous. Family TESTACELLIDAE Shell small, auriform, situated at the posterior end of the body, covering the pallial organs. The upper tentacles (ommatophores) are not bulbous at the apex. There is no caudal mucus gland. A pair of lateral grooves with side branches directed obliquely upwards and forwards, and downwards and forwards, arises from a peri- pallial groove. The pedal mucus gland lies free in the body cavity. The cerebral commissure is very short, and the cerebro-pleural and cerebro-pedal commissures long. There is no jaw. The buccal bulb is enormous. The long narrow radular teeth, barbed at the apex, are disposed in transverse V-shaped rows, and a central tooth may be absent. The stomach is feebly developed and the anus lies just within the pulmonary cavity posteriorly. The kidney is without a secondary ureter. An osphradium is said to be present in a groove in the floor of the mantle cavity (Plate, 1891). The vagina is without accessory glands, and the right ocular retractor muscle passes forwards between the penis and vagina. Testacella is carnivorous, feeding on earth-worms and occasionally on other slugs, and is largely subterranean in habit. Genus TESTACELLA Draparnaud, 1801 Type species Testacella haliotidea Draparnaud, 1801. The sole genus, with the characters of the family. Through adaptation to a life spent largely underground, and a diet of earth-worms, the structure of Testacella has been profoundly modified. The relatively large shell and mantle cavity of other snails have been greatly reduced in size, and the body anterior to these structures elongated to accommodate the enormous buccal bulb with its powerful muscles and barbed radular teeth. The backwards shift of the pallial region is accompanied by a clockwise rotation of 180, so that the heart and pericardium lie on the right side of the kidney with the ventricle in front of the auricle. The normal backwardly- directed loop of the intestine is obliterated, and the rectum runs backwards instead of forwards to the anus (beside the pulmonary orifice, at the rear). It is convenient to gather together the available information on the life history of all the three British species of Testacella. A recent paper by Stokes (1958) gives further information, supplementing that of Barnes & Stokes (1951). Mating Little is known, but Gerhart (1940 : 579) reports that while unpacking a container he found a pair of T, scutulum in coitus in the moss, and accidentally separated them. BRITISH SLUGS 109 The penes were bluish-white, translucent, elongated, dorso-ventrally flattened cones, withdrawn from the adjacent vaginal openings. No appendix was visible, and the penes were of the simplest type occurring in land pulmonates. They quickly disappeared within the owners' bodies, and no more was to be seen. Barnes & Stokes (1951) found that the coitus of T. haliotidea and T. scutulum took place under the surface of the soil in their containers, the animals lying head to head : they do not mention any visible everted organs. Chemin (1939^ and b) reports self-fertilization in T. haliotidea isolated when young. Development The eggs are ellipsoidal, with a white calcareous shell which becomes yellow, and are laid underground. Those of the three British species are distinguished only by differences in size 15x4 mm for T. maugei, 7x4 mm for T. haliotidea, and 4x3 mm for T. scutulum. Barnes & Stokes confirm the old statement that they explode when placed on a warm hand. The eggs of T. maugei are laid in groups of 8-15 during the summer months, those of T. haliotidea and T. scutulum in captivity (Barnes & Stokes, 1951) in batches of 2-14 in March, 54-95 days after mating. Taylor (1902-07) records the time of development as 30-35 days for T. maugei, 20-36 days for T. scutulum and (quoting Gassies, 1849) 10-22 days for T. haliotidea. However, Barnes & Stokes found that embryonic development takes much longer than this for the latter two species : 144-201 days, or five to six months. Like all our slugs and some, perhaps all, of the shell-bearing Stylommatophora except Succinea (which lacks a caudal vesicle), Testacella exhibits in the middle and late stages of development a haemocoelic cephalic vesicle protruding beneath the mantle over the head, containing part of the viscera ; and a caudal haemocoelic vesicle projecting beyond the tail, which contains only haemocoelic fluid and cor- puscles. The caudal vesicle varies in shape from genus to genus (see Text-fig. 18). In Testacella it is a flattened bilobed or broadly cordate sac (Text-figs, i D and 18 c). The two vesicles pulsate, expanding and contracting alternately so that haemocoelic fluid is driven back and forth, at a rate depending on the temperature. The vesicles may have an excretory or a respiratory function. The cephalic vesicle is withdrawn with its contents into the body, and shortly before hatching the caudal vesicle also shrinks and disappears. In Testacella, with its posterior mantle, the cephalic vesicle does not extend over the head (at any rate in the later embryonic stages). Behaviour It has been noted (Adams, 1941) that dust sprinkled on Testacella leads to copious discharge of fluid which washes it away. Another phenomenon does not seem to have been recorded. When lightly irritated, Testacella maugei partially contracts, and sometimes a forcible discharge of air and froth occurs in the middle line from under the front of the mantle edge, with a distinct high-pitched hissing sound. Presumably a communication to the lung cavity exists, and air is expelled by closure of the pulmonary orifice and muscular contraction. No duct is apparent on gross dissection, but serial section of the area show spongy tissue leading to the anterior no BRITISH SLUGS expansions of the pulmonary cavity, through which air might be expelled. In its underground travels, or if dragged into a burrow by an attacked earth-worm, it is easy to imagine that earth could become impacted under the mantle border, and that the forcible expulsion of air and fluid would serve to dislodge it. Ecology Mostly to be found in gardens, allotments and other cultivated ground. During the day Testacella remains underground, retreating deeper in dry weather but some- times hiding under stones and herbage in damp weather ; while at night it crawls freely on the surface. In the winter it hibernates in a strongly contracted state, in an underground cell. On slowly approaching a worm, the slug shoots out its radula and impales its prey at one end or in the middle, and engulfs it within a few minutes. Another worm may be eaten after a day or two. The faeces are very small in bulk and contain discarded radular teeth, fragments of worm cuticle and setae, and sometimes small white bodies which may be from the calcareous oesophageal glands of worms. Testacella is also said to eat other slugs. Testacella maugei Ferussac, 1819 Testacella maugei Ferussac, 1819 : 94, pi. 8, figs. 10 and 12. Teneriffe. Testacella aurigaster Watson, 1915 : 221 (from Layard MS.). Notes The type was found in 1796 by M. Mauge ; who died before publishing an account. According to Turton (1831 : 27), it was sent in 1812 from Dr. Leach at Bristol to Ferussac. T. aurigaster refers to this species, introduced into South Africa (Watson, 1915 : 221). First British record By Ferussac (1819). Diagnostic features Pear-shaped appearance (body widest behind) ; shell large (14 X 7 mm) ; two median lines of dorsal tubercles conspicuous, especially in the young ; lateral grooves wide (5 mm) apart at origin. Buccal bulb lacks lateral retractors ; cephalic retractors arise asymmetrically, from left side. Penis swollen posteriorly, lacks the flagellum of T. haliotidea ; penial retractor arises from body-wall about two- thirds of the way back ; vas deferens convoluted near free oviduct ; spermatheca duct long and slender, swollen proximally. External appearance (PI. I, fig. 13) Length when extended, 6-10 cm. The most bulky part of the body lies behind the centre. The lateral grooves are separated by about 5 mm at their origin from the peri-pallial groove, and each gives rise to about fourteen dorsal and ventral BRITISH SLUGS in grooves directed obliquely forwards, which subdivide to produce the reticulate surface pattern. The row of tubercles on each side of the median dorsal line is more prominent than the other dorsal rows, especially in the young. The upper tentacles, bearing the small black eyes, are not bulbous at the apex, and the lower tentacles are short. The true pedal groove is only distinct for a few millimetres at the front of the foot fringe, but a peripodial groove above this runs to the tail. There is no caudal mucus gland. When crawling on a glass plate, the yellow sole is seen to be obscurely tripartite. The mucus is clear and viscid, drawing out into thin threads. The body is usually some shade of brown more or less sprinkled with darker spots, paler on the sides, and the sole is yellow or pink. The mantle is speckled with black, and the paler margin slightly overlaps the shell. The posterior respiratory orifice is bounded by the right and left mantle lobes. Shell (Text-fig, i A) The shell measures about 14 X 7 mm, and has a brown periostracum, often abraded. It is more convex than the shells of the other two species, and the columella does not show the same callous enlargement at the hinder end. The crescentic muscle scar is visible on the under surface. External variation Variety aperta Taylor (1902) has the shell wider and flatter than normal. Varieties albina, griseonigrescens, viridans and griseorubescens Gassies & Fischer (1856), nigra Collinge (1898) and aurea Taylor (1902) are colour varieties, whose charac- teristics are suggested by their names. P allied organs (Text-figs. 2 D and F) The respiratory cavity, bounded by the right and left mantle lobes, leads to the pulmonary chamber, which is bounded by spongy vascular tissue. Anterior to this, the chamber is prolonged forwards as thin -walled air sacs projecting into the visceral cavity. The somewhat quadrate kidney, opening without a secondary ureter at the back of the chamber, lies above and in front of the pulmonary chamber. The pericardium lies to the right of the kidney and the ventricle is anterior to the auricle, and both branches of the aorta run forwards. The rectum runs obliquely in the pulmonary chamber to terminate in the anus just above and on the left of the respiratory orifice. The kidney projects over the apex of the left liver lobe. An osphradium is represented by a ridge of cells in the floor of the respiratory chamber, and is innervated from the right visceral ganglion (Plate, 1891). Retractor muscles (Text-fig, i E) Both the cephalic retractor muscles arise from the body-wall on the left side. The right ocular retractor passes between the penis and the vagina, and a branch takes the same course to the side of the head. The slender penial retractor arises from the right body wall about two-thirds of the way back, and not from the posterior end as it does in T. haliotidea and T. scutulum. 112 BRITISH SLUGS FIG. i. Testacella. Reproductive organs (x 3-3), shell (x 2). A. T. maugei shell below. B. T. haliotidea shell right, below. C. T. scutulum shell right, above. D. T. scutulum advanced embryo ( x 6-7) showing cephalic vesicle in front of shell and caudal vesicle posteriorly. E. T, maugei buccal bulb and ocular retractors. BRITISH SLUGS Nervous system (Text-fig. 2 E) The cerebral commissure is very short, and the cerebro-pleural and cerebro-pedal commissures are long to embrace the large buccal bulb. The buccal ganglia are large. The five visceral ganglia are close together, but not fused. Radula (Text-fig. 2 A) In each row, sixteen aculeate barbed teeth on each side of a small median tooth increase in size up to the penultimate one, while the sixteenth is much smaller. There are about fifty V-shaped rows. B FIG. 2. Testacella. Representative radular teeth ( x 67). A. T. maugei showing central tooth. B. T. haliotidea showing absence of central tooth. C. T. scu- tulum showing absence of central tooth. T. maugei. D. Posterior end of body ( x 3-3) with shell removed and right mantle lobe reflected to expose respiratory orifice. E. Central nervous system, showing buccal ganglia (black) posterior to cerebral ganglia when animal is contracted, F. Pallial region, viewed as if shell were transparent. H4 BRITISH SLUGS Alimentary system The buccal bulb, which is large but does not extend as far back as in T. haliotidea and T. scutulum, lacks lateral muscles but has a bundle of from two to four pairs of powerful muscles at the posterior end, attached to the diaphragm towards the left of the mid-line (Text-fig. I E). The small stomach is scarcely wider than the oesophagus and intestine. It receives the right and left liver ducts, and the left liver lobe forms the apex of the visceral mass. The intestine makes one forwardly- directed loop, and the long rectum runs backwards to the anus, close to the respiratory orifice. Reproductive system (Text-fig. I A) The reproductive organs are simple. The pale ovotestis lies at the apex of the visceral mass between the right and left lobes of the liver. The slightly sinuous hermaphrodite duct becomes slender and folded before entering the long narrow albumen gland, and bears a small seminal vesicle. The convoluted vas deferens, arising from the yellow ribbon-like prostate, runs forwards beside the free oviduct and then turns backwards, beneath the right ocular retractor to the apex of the penis. The penis is about 9 mm long, slender in front and dilated posteriorly and often folded over forwards. The mucous membrane is papillate at the apex and longitudinally ridged beyond, and there is no penial papilla. There is no caecum or flagellum. The penial retractor, arising from the right body wall about two- thirds of the way back, is inserted at the apex of the penis. The atrium is short and opens below the right upper tentacle. The free oviduct and vagina are wider than the penis. The long slender spermatheca duct is dilated at its origin from the oviduct, and enters the spermatheca sac laterally. The mucous membrane of the duct is longitudinally ridged, the ridges being more prominent in the bulbous part of the duct. Mating, Development, Behaviour, Ecology See pp. 108-110. The eggs measure 5x4 mm. Distribution (Map i) T. maugei is the most westerly of the three British species of Testacella, known from Tangier, south-western Spain, Portugal, Brittany, Dieppe and the Atlantic isles (and introduced into North America and South Africa) . In England and Wales it occurs west of a line from Cheshire to Middlesex, and in Ireland south-east of a line from Dublin to Cork. It is probably truly native only in the south-west. Fossil record Holocene and Recent. Material examined Living specimens from Singleton Park Botanical Gardens, Swansea, Glamorgan. Spirit material from various British localities, the Azores and Madeira. BRITISH SLUGS 115 Testacella haliotidea Draparnaud, 1801 Testacella haliotidea Draparnaud, 1801 : 99. ? Montpellier. Testacella europea de Roissy, 1805, 5 : 252. First British record By Lukis (1834), fr m Guernsey. Diagnostic features Less bulky posteriorly, when extended, than T. maugei ; shell much smaller, (7x5 mm), columella somewhat truncate anteriorly, with conspicuous callous enlargement posteriorly ; dorsal tubercles not especially prominent ; lateral grooves close together at origin. Buccal bulb with powerful lateral retractors ; cephalic retractors arise symmetrically. Radula lacks a central tooth, apophyses of marginal teeth nearly central. Penis with flagellum and short apical caecum ; penial retractor arises from diaphragm at posterior end of body ; spermatheca duct short and bulbous. External appearance (PI. i, fig. 10) Length when extended, 8-12 cm. The lateral grooves are close together at their origin from the peri-pallial groove. Colour usually dull greyish-yellow, with a pale sole and foot fringe. Shell (Text-fig, i B) Measures about 7x5 mm, convexly auriform, triangular ; brown periostracum usually much abraded. Columella usually slightly truncate anteriorly, and expanded to a rounded callus posteriorly. External variation In var. trigona Gassies & Fischer (1856) the shell is thicker and the right margin expanded, in var. major G. & F. the shell is larger (11-5 X 7 mm). The body is yellow in var. flavescens Moquin-Tandon (1855), and whitish in var. albino, Moquin- Tandon. Pallial organs As in T. maugei. Retractor muscles The cephalic retractor muscles arise symmetrically on each side of the body-wall about two-thirds of the way back, and not from the left side only as in T. maugei. The long, slender penial retractor arises dorsally at the hind end of the body. Nervous system > As in T. maugei. n6 BRITISH SLUGS Radula (Text-fig. 2 B) The median tooth is vestigial or absent. On each side, eighteen barbed aculeate teeth gradually enlarge up to the penultimate one, while the eighteenth is shorter. The lateral apophysis is nearly midway between base and apex (especially towards the margin of the radula), while in T. scutulum it is nearer the basal end. There are about fifty V-shaped rows. Alimentary system The large buccal bulb is longer than in T. maugei and has, in addition to the bundle of posterior muscles attached to the diaphragm, a series of ten to fifteen pairs of lateral muscles attached to the posterior half of the bulb and arising from the left body-wall. The relations of the liver lobes, intestine and rectum are as in T. maugei. Reproductive system (Text-fig, i B) Compared with T. maugei, the central portion of the hermaphrodite duct is longer and more folded, the penis is slighly enlarged at the entrance of the vas deferens, and opposite this point there is a short blunt caecum. There is also a flagellum, which varies in length but may be two-thirds as long as the penis, from the apex of which a long slender retractor muscle runs to the diaphragm at the posterior end of the body. There is no penial papilla, and the mucous membrane is papillate. The spermatheca duct is short, wide and bulbous with the mucous membrane strongly folded longitudinally, and the sac is oval. Mating, Development, Behaviour, Ecology See pp. 108-110. The eggs measure 7x4 mm, and so are considerably larger than those of T. scutulum. Distribution (Map 2) T. haliotidea is the most widely distributed species, ranging from the Atlantic isles to the Balkans, and from the North African coast to south Scotland. It probably occurs in every county in Britain up to Stirling, and in the south-eastern and southern counties of Ireland. Fossil record Holocene and Recent. Material examined Living specimens from Henley-on-Thames, Bedford and Cambridge. Spirit material from Britain, France and Majorca. BRITISH SLUGS 117 Testacella scutulum Sowerby, 1820 Testacella scutulum Sowerby, 1820 : pi. 159, figs. 3-6. Kennington Road, Lambeth, Surrey. Notes Taylor (1902-07) considers T. bisulcata Risso (1826) as a variety of this species. Plate (1891) and Hoffman (1925), however, consider it to be quite distinct. Plate's figure, copied by Hoffman, shows a flagellum from the anterior end of the penis, and a tubular connection from the vagina to the apex of the flagellum. This most unusual arrangement is said to be confirmed by serial sections. First British record By Sowerby (1820). Diagnostic features Resembles T. haliotidea, but differs in the following. Shell smaller (6 X 3-5 mm) and flatter ; animal yellow in colour, speckled with brown. Apophyses of marginal radular teeth more basal. Penis lacks caecum and flagellum ; free oviduct shorter than vagina ; spermatheca duct longer and narrower (but quite different from that of T. maugei). External appearance Length when extended, 8-12 cm. The lateral grooves are said to be confluent for about i mm at their origin from the peri-pallial groove, but this does not seem to be constant. Colour some shade of yellow, more or less sprinkled with brown ; sole and foot fringe brighter yellow. Shell (Text-fig, i c) Smaller than that of T. haliotidea (6 X 3-5 mm), and flatter sometimes even concave. The brown periostracum tends to be retained better than in the other species. The columella is slightly truncate anteriorly and expanded posteriorly, as in T. haliotidea. Young shells of these two species can easily be confused. External variation Most of the varieties mentioned by Taylor (1902-07) refer to continental forms, which may or may not be conspecific with T. scutulum. However, var. aurea Cockerell (1885) presumably refers to this species : it is orange-coloured, with brown mottling and a bright orange foot sole. Pallial organs As in T. maugei and T. haliotidea. n8 BRITISH SLUGS Retractor muscles The cephalic retractors arise symmetrically from the body-wall about half-way back. The penial retractor arises far back, from the diaphragm. Nervous system As in T. maugei and T. haliotidea. Radula (Text-fig. 2 c) As in T. haliotidea, the median tooth is absent and there are eighteen teeth on each side of the mid-line ; but in the outer marginal teeth the apophyses are nearer the bases. There are about forty-five V-shaped rows. Alimentary system The buccal bulb has three or four strong muscles from the diaphragm to its posterior end, and the hinder half has five to ten pairs of lateral muscles attached to the left body-wall. The relations of the liver lobes, intestine and rectum are as in T. maugei and T. haliotidea. Reproductive system (Text-fig. I c) The ovotestis, albumen gland, prostate and oviduct resemble those of T. haliotidea, but the free oviduct is relatively shorter and the vagina longer. The spermatheca duct is intermediate in length and breadth between those of T. haliotidea and T. maugei. .The penis is cylindrical, without dilatations or flagellum, and the retractor muscle arises far back from the diaphragm. The penial mucous membrane is papillate, and that of the spermatheca thrown into longitudinal folds. Mating, Development, Behaviour, Ecology See pp. 108-110. The eggs are smaller (4x3 mm) than those of the other species, but otherwise similar. Distribution (Map 3) Taylor (1902-07 : 18) apparently treated T. bisulcata Risso (1826) as a variety of T. scutulum, so that his account of the latter's continental distribution is unreliable. In the British Isles, its distribution is almost the same as that of T. haliotidea. Fossil record Recent. Material examined Living specimens from Bedford. Spirit material from Britain and Corsica. BRITISH SLUGS 119 Superfamily ENDODONTACEA Shell spiral depressed, or reduced to a flat plate or mass of granules enclosed by the mantle. A supra-pedal groove is present, and a caudal mucus gland may be present or absent. The jaw is either crescentic and ribbed, or plaited. Marginal radular teeth generally aculeate. The genitalia are simple. Family ARIONIDAE Slug-like animals with the mantle covering only the anterior part of the body, with the respiratory orifice in a slit on the right margin. Shell an oval plate or mass of granules except in Binneya, where it is exposed and spiral. The sole is uniform or tripartite. The kidney surrounds the pericardium. An epiphallus is present. The jaw is ribbed (odontognath) , the central radular tooth tricuspid, lateral teeth bicuspid, marginals with short wide basal plates, with long mesocones and with or without ectocones. The Arionidae require relatively high humidity, and dry barren areas are barriers to their dispersal. The family seems to have been derived from Endodontid stock with spiral shells, the cephalic and buccal retractor muscles with a common origin from the columella, and a penial papilla. In the primitive North American genus Binneya, the visceral hump is capped by a spiral Vitrina-like shell, the tail is solid and the retractors have a common origin. Various genera show transitions from this snail-like condition to that of the most modified subfamily, the Arioninae of the western Palaearctic Region. For a discussion of the probable evolution of the other subfamilies and genera, see Pilsbry (1948 : 660-665). Subfamily ARIONINAE Slugs with a wide foot fringe ; caudal mucus gland present ; the shell is reduced to a flattened plate in Geomalacus, and to discrete granules in Arion. The cephalic retractor muscles are widely separated at their origin from the posterior border of the diaphragm. The stomach apex lies behind the posterior loop of the intestine. The ovotestis lies behind all the intestinal loops. There is no penis or vagina, and the oviduct, spermatheca and epiphallus enter the upper atrium separately. Of the British genera, Geomalacus with its solid shell, anterior reproductive orifice, narrower foot fringe, and extreme western distribution, seems to be the more primitive. Genus GEOMALACUS Allman, 1843 Type species Geomalacus maculosus Allman, 1843. Genital orifice below and behind the right upper tentacle ; atrium greatly elong- ated ; genital retractor muscle arising from the dorsal body-wall at the extreme posterior end of the body and inserted into the spermatheca duct ; shell a compact oval calcareous mass. ZOOL. 6, 3. 9 BRITISH SLUGS G. maculosus is the only member of this genus occurring in the British Isles, though one or two more are found with it in the Iberian peninsula. Geomalacus maculosus Allman, 1843 Geomalacus maculosus Allman, 1843 : 851. Lough Caragh, Co. Kerry, Ireland. Notes The types were found by W. Andrews of Dublin in 1842. First British record By Allman (1843). Diagnostic features Distinguished from species of Arion, externally by the anterior situation of the reproductive orifice, narrower median area of the sole, and maculate colour pattern ; and internally by the solid oval shell, greatly elongated atrium, and long genital retractor arising far back from the mid-dorsal line of the body-wall. External appearance (PL i, fig. i) Length when extended about 5-5 cm, but occasionally up to 9 cm. The granulose mantle occupies about one-third of the body, with the respiratory orifice a little in front of the centre of the right border. The reproductive orifice is nearer to the right upper tentacle than to the respiratory orifice. The supra-pedal groove is deep and distinct and the pedal groove indistinct. The caudal mucus gland is a transverse slit. The body shows about twenty-five rows of polygonal tubercles (Text-fig. 3 c). The colour is grey, brown or nearly black, irregularly marked with yellow or white maculations more or less developed. The young have two longitudinal bands on each side, which tend to become obscured in the adult, but sometimes remain. In contrast to Arion, Geomalacus sometimes rolls up like the woodlouse Armadillidium. External variation Var. fasciata Cockerell (1892) retains the banding characteristic of the young animal white or whitish with black or brown markings : mantle with dark marbling and lateral bands, body with a dark subdorsal band along each side. Charles Oldham bred in captivity a crimson form, var. rutilans Oldham (19426), which has not been observed in the wild. The other varieties depend merely on the relative development of dark pigment on a pale ground : var. allmani Heynemann (1873), dark brown or grey with whitish maculations ; var. verkruzeni Heynemann, grey, darker dorsally, with whitish maculations; var. andrewsi Mabille (1867), whitish, overspread with blackish spots. BRITISH SLUGS 121 Shell (Text-fig. 3 i) Variable, but usually solid, oval, convex above and concave below, measuring about 4x3 mm. PaLlial organs (Text-fig. 3 B) The kidney completely surrounds the pericardium, and the cavity is occupied by lamelliform folds. The ventricle lies behind the auricle. The anterior branch of the aorta follows the usual course, arching over and passing beneath the anterior intestinal loop, and running forwards between the visceral and pedal ganglia to supply the anterior end of the body. The posterior aorta runs backwards to the digestive and genital systems. Retractor muscles (Text-fig. 3 B) The right and left cephalic retractor muscles are widely separated at their origin from the diaphragm, behind the kidney. The buccal retractor arises from the dorsal body-wall, a little further back and slightly to the right of the middle line. The right tentacular branch lies to the left of the oviduct and epiphallus, and does not pass between them. Nervous system The cerebral commissure is of medium length. The cerebral ganglia give rise to nerves to the head and lips, the peritentacular nerves, the combined ocular and olfactory nerve, the nerve to the lower tentacle, and on the right side to the genital nerve supplying the epiphallus and vas deferens. The right pleural, visceral and abdominal ganglia are fused together, but just distinguishable, and the same is true of the left pleural and the left visceral ganglia. The pleuro-pedal commissures are very short. The buccal ganglia, as well as supplying the buccal bulb, send a nerve to the salivary glands. Jaw (Text-fig. 3 D) The jaw is about 3 mm across, dark brown, arcuate, and crossed by eight or ten broad ribs which slightly denticulate the upper and lower margins. The number of ribs may be reduced by fusion, and there is sometimes a slight median projection. Radula (Text-fig. 3 E) The radula is about 8 mm long. There are usually about 200 rows of teeth, but the number varies considerably. The central tooth has very small ectocones. The lateral teeth, about twenty in number, have a large mesocone, a small ectocone, and no endocone. The ectocone gets smaller in the outer lateral teeth. In the marginal teeth it descends to the base of the mesocone, and gets larger again as the margin of the radula is approached. 122 BRITISH SLUGS 52. H FIG. 3. Geomalacus maculosus. A. Reproductive organs (x 2-7). B. Head and mantle region (x 1-7) with mantle reflected to the right showing kidney, peri- cardium, rectum, terminations of genital and alimentary systems, origin of pharyngeal and left ocular retractors cut short. C. Polygonal body-tubercles. D. Jaw (x 13). E. Representative radular teeth ( x 330). I. Shell. Avion. Representative radular teeth (x 330). F. A. ater. G. A. lusi- tanicus. H. A. subfuscus. BRITISH SLUGS 123 Alimentary system The oesophagus, after passing through the nerve ring, enlarges to a crop, merging into the stomach which forms the most posterior part of the alimentary canal. The narrower intestine passes far forwards (to the level of the kidney) and then turns back, but does not reach the stomach before it runs forwards again, as the rectum, to end in the anus in the posterior lip of the respiratory orifice. The short ducts of the anterior and posterior lobes of the liver open at the junction of stomach and intestine. Reproductive system (Text-fig. 3 A) The somewhat raspberry-shaped brown ovotestis lies between the upper and lower lobes of the liver, at the apex of the stomach. The pale hermaphrodite duct passes forwards, becoming wider and folded, towards the linguiform albumen gland. It then becomes narrower, runs backwards between the lobes of the albumen gland, and turns sharply forwards, forming a " claw" or seminal vesicle. The common duct is much folded, the prostatic portion showing as a yellow ribbon. The narrower free oviduct enters the lower end of the atrium ; its mucous membrane is thrown into fine longitudinal folds. The lower atrium shows internally a complicated mass of large sinuous folds, and is also remarkable in having a long slender diverticulum with circular folds, extending backwards as far as the albumen gland. The short duct of the globular spermatheca enters the apex of this diverticulum. The vas deferens runs backwards from the free oviduct towards the apex of the atrial diverti- culum, to enter the wider epiphallus which is very long and much coiled, the coils being bound together by connective tissue. The epiphallus enters the atrial diverti- culum close to the spermatheca duct. A long slender retractor muscle, arising from the mid-dorsal line at the posterior end of the body, is inserted at the neck of the spermatheca and apex of the diverticulum. The mucous membrane of the epiphallus shows longitudinal rows of papillae, and that of the spermatheca longitudinal folds. A membrane with a series of muscular bands extends from the atrial diverti- culum to the diaphragm, and similar bands from the lower atrium to the body-wall. Mating This has not been recorded. Development The ellipsoidal eggs are said (Taylor, 1902-07 : 256) to be relatively large, from 8-5 x 4-25 to 6 X 4 mm, and to be white when laid, turning brown. They are laid, in clusters of from 18 to 24, in July and August, and hatch in from six to eight weeks. The young, which mature in the following summer, show two dark bands on each side which fade during growth. Behaviour When irritated, it can roll up like the woodlouse Armadillidium (Taylor, 1902-07 : 257)- 124 BRITISH SLUGS Ecology G. maculosus lives on the bare mountain slopes of Old Red Sandstone in Cork and Kerry, where it browses on the algae, mosses and liverworts which coat the boulders. In captivity it readily feeds on lettuce and carrot, and has been observed to attack small snails such as Vitrina pdlucida. Distribution (Map 4) Portugal, north-western Spain, and Vennes (France). In Britain it is confined to south-western Ireland, in the counties of Cork and Kerry. Fossil record Recent. Material examined Preserved specimens from Glengariff, Co. Cork (August 1931), and other localities, in Cork and Kerry. Genus ARION Ferussac, 1819 Type species Avion empiricorum Ferussac (= Limax ater Linnaeus, 1758). The diagnostic features of the genus are those of the subfamily Arioninae. Differences from Geomalacus are : reproductive orifice immediately below posterior lip of respiratory orifice ; upper atrium not prolonged backwards as a diverticulum ; genital retractor short, arising from diaphragm close to left cephalic retractor, and inserted into both spermatheca duct and free oviduct ; shell reduced to discrete granules ; median area of sole wider. A . intermedius and A . fasciatus seem to be less advanced than the other species, in the absence of a ligula, the limited eversion of the gentalia during coitus (in A. fasciatus at least), and the slightly more anterior position of the reproductive orifice in A . fasciatus. A . hortensis differs strikingly from A . fasciatus and from the larger species, but agrees with A . subfuscus and A . lusitanicus, in having the ligula (or its functional equivalent) in the proximal part of the oviduct. A. lusitanicus is in some respects intermediate between A . subfuscus and A . ater ; while the latter, in which the banding pattern disappears completely in the adults, is probably the least primitive. The various subgenera and sections that have been proposed are not satisfactorily delimited. Hesse (1926 : 64-69) uses the following subgenera : Lochea Moquin- Tandon (1855), for A. ater and A. lusitanicus (though the genitalia of the latter are much more like those of A . subfuscus) ; Mesarion Hesse, for A . subfuscus ; Carinarion Hesse, for A. fasciatus ; Kobeltia Siebert (1873), for A. hortensis ; and Microarion Hesse, for A . intermedius. BRITISH SLUGS 125 Arion intermedius Normand, 1852 Avion intermedius Normand, 1852 : 6. Valenciennes. Arion minimus Simroth, 1885 : 237, pi. 7, fig. 41. First British record By Alder (1848 : 124), as Arion flavus. Diagnostic features Small size, tubercles assume a conical shape with translucent apices when animal contracts, yellow mucus accumulates at each end of foot ; reproductive orifice slightly anterior to respiratory orifice. External appearance (PL i, fig. 7) This is the smallest British Arionid, about 2 cm long when crawling. Its colour varies from yellow to some shade of grey, with the head darker grey or black, and there may or may not be body and mantle bands. All these colour varieties may be found living together. A. intermedius is easily distinguished from the young of larger species by the form of its tubercles, which stand up in little conical eminences with translucent tips when the animal contracts hence the popular name Hedgehog Slug. The pair of mantle bands together form a lyre-shaped figure, and the right band arches over the respiratory orifice and sends a branch down below and behind it. The sides of the body are paler than the back, the foot fringe is yellowish-grey, and just above it there is a row of black dots towards the anterior end. The sole is yellowish-grey, and yellow mucus tends to accumulate at each end of it. The genital orifice is immediately below the slit of the respiratory orifice, and thus is further back than in A . fasciatus but slightly further forward than in the other species. External variation Var. normalis Moquin-Tandon (1855), yellow with blackish head and tentacles ; var. plumbea Collinge (18926), dark grey ; var. brunnea Taylor (1902-07), maroon- brown with darker bands ; var. alba Taylor, almost white. Shell The shell granules are not specifically distinguishable. In calcareous sites they may be aggregated into an irregular mass. Pallial organs As in A. ater (p. 139). Retractor muscles As in A. ater, except that the cephalic retractors do not arise at the same level, but with the right a little further forward than the left ; while the genital retractor 126 BRITISH SLUGS sends a branch to the junction of oviduct and atrium, as well as to the neck of the spermatheca. Nervous system As in A . ater. Jaw About i mm across, arcuate and crossed by seven to ten widely-spaced ribs. Of the usual Arionid shape, and so not like Taylor's figure (1902-07, fig. 248). Radula About 2 mm long, with about 120 transverse rows of teeth. The base of the mesocone of the central tooth is more constricted than in the other species. There are about sixteen lateral and sixteen marginal teeth, each with mesocone and ectocone : they do not have endocones. Formula C.i6.i6 x 120. Alimentary system As in A . ater, except that the visceral mass is less twisted. Reproductive system (Text-fig. 4 H) The deeply pigmented ovotestis lies between the lobes of the liver, and is partly exposed on the surface of the visceral mass. The central part of the pale hermaphro- dite duct is wider and somewhat sinuous, then narrows and runs back on the short linguiform albumen gland and bends sharply forwards, making a " claw " or seminal vesicle. The sperm-oviduct is short, and the free oviduct enters the very short upper atrium without a ligula. The mucous membrane is thrown into five or six low longitudinal folds. The lower atrium is about 2 mm long and smooth internally. The spherical spermatheca is connected to the upper atrium by a very short duct, much dilated at its base. The vas deferens is about 2 mm long, and the epiphallus about the same length. Spermatophore Unknown. Mating Not recorded. Development The ellipsoidal eggs are relatively large (2 X 1-5 to 2 X 2 mm), opaque, white and dotted with calcareous spots. They are laid in the summer and autumn, and hatch in about three weeks. The young at birth are 4 mm long, and easily recogniz- able as miniatures of the adults. BRITISH SLUGS 127 Ecology Common in the moist ground-litter of woods and copses, beneath logs, and at the roots of grass in open fields. Often found feeding on toadstools in the wild, and in captivity readily eats carrot, lettuce and oatmeal. Distribution (Map 5) Portugal, Spain, France, northern Italy, Switzerland, Germany, Scandinavia, Russia, Azores (and introduced into New Zealand, Polynesia and North America). It probably occurs in every vice-county in the British Isles, including islands north- wards to the Orkneys and Shetland. Fossil record Although shell granules of Arion are not recognizable with certainty, they are recorded from the Cromerian onwards. They cannot be assigned specifically. Material examined Living specimens from many parts of England and Wales. Spirit material from Britain, France, Norway and Tahiti. Arion fasciatus (Nilsson, 1822) Limax fasciatus Nilsson, 1822 : 4. ? Sweden. Arion circumscriptus Johnston, 1828 : 76. Notes Collinge (18926) reinstated Nilsson's name, though it has not been adopted by other British authors. Lohmander (1937) also concluded that fasciatus is the valid name, though he continued to use circumscriptus. Nilsson's description of his species leaves no doubt that it is the A. circumscriptus of most authors, and Collinge's use of fasciatus prevents any action being taken to conserve circumscriptus. First British record By Johnston (1828). Diagnostic features Body bell-shaped in transverse section when contracted, mid-dorsal line of tubercles pale and raised, reproductive orifice in front of respiratory orifice, right mantle band arches over respiratory orifice, sole opaque porcelain white ; sperma- theca with short stout duct and long tapered sac, spermatophore coiled anteriorly. External appearance (PI. I, fig. 5) Length when crawling about 3-3-5 cm. Easily distinguished from A. hortensis by the opaque white sole, and by the bell-shaped transverse section of the body 128 BRITISH SLUGS when contracted (caused by splaying-out of the lower part). The reproductive orifice is further forward than in our other species, being a little in front of the slit of the respiratory orifice. Body and mantle some shade of grey, becoming paler towards the foot and with a dark lateral band on each side. The mantle bands form a lyre-shaped figure, and the right band arches over the respiratory orifice without sending a branch below it. The bands are set at a higher level than in A . hortensis. The tubercles of the mid-dorsal line are larger and paler than those on each side, forming a keel (hence the sectional name Carinarion Hesse, applied to this species) : this is often more distinct in the young. The foot fringe is somewhat lineolated, especially towards the tail. External variation Var. grisea Collinge (i8Q2c), light silver-grey with dark bands; var. flavescens Collinge (1892^) (= var. circumscripta 0kland, 1923), light brownish-yellow, with dark bands and a yellow zone over the foot. Lohmander (1937) discusses whether these and other named forms are varieties or subspecies. Pallial organs As in A . ater. Retractor muscles The genital retractor is mainly inserted into the duct of the spermatheca, but usually sends a small branch into the middle of the free oviduct. Otherwise as in A. ater. Nervous system As in A . ater. Jaw Of the usual arcuate shape, crossed by ten to twelve ribs ; about 1-25 mm across, coloured yellow to pale brown. Radula Closely resembles that of A. hortensis, but the marginals often have the small ectocone better developed. Length 3-4 mm, formula .11.23 X 120. Alimentary system As in A . ater, except that the visceral mass is less twisted. Reproductive system (Text-fig. 4 i) The ovotestis is darkly pigmented, the hermaphrodite duct has the usual " claw " or seminal vesicle at its entry into the albumen gland, and the sperm-oviduct is pinkish in colour, with a broad yellow prostatic ribbon. The free oviduct is rather BRITISH SLUGS 129 long and slender, agreeing with that of A . intermedium in not being dilated terminally to contain a ligula as in A . hortensis. The spermatheca has a short stout duct and an elongated sac, wide at the base and tapering apically. It varies a good deal, sometimes assuming bizarre shapes. The vas deferens is about the same length as the epiphallus, which terminates with a slightly marked annular swelling. The upper atrium is very small, and contains no ligula. The lower atrium is long, dorso- ventrally flattened, and yellow. Spermatophore (Text-fig. 4 G) A chitinous tube about 4 mm long, with five low ridges of which the most prominent is finely toothed. Within the spermatheca, the Spermatophore is bent into a U, and is swollen for a short distance behind the slender coiled anterior end. On this swollen part, Lohmander (1937) describes, but does not figure, a small projection which anchors the Spermatophore to the mucous membrane of the spermatheca. 0kland (1923) figures an incomplete Spermatophore, and Lohmander (1937) complete ones. Mating Coitus lasts about twenty minutes. In contrast to A. hortensis, A. subfuscus and A . ater, there is little or no eversion of the genitalia. Development The ellipsoidal eggs, 3x2 mm, are yellow or sometimes dark amber, more flexible and translucent than those of other species of A rion. They are laid in clutches of up to thirty, from May to the autumn, and hatch in four or five weeks (or longer in cold weather). The newly-hatched young are 5 mm long, pale yellowish-grey ; not darker dorsally, but with a pale central line of larger tubercles forming a keel towards the tail. The lyre-shaped figure formed by the mantle bands is faint in front and darker posteriorly. Ecology A . fasciatus is common amongst dead leaves and moss and under logs and stones, in thickets, fields and waste places, and sometimes in gardens. Distribution (Map 6) Probably every vice-county of the British Isles, and in Iceland (? introduced), the Faroes, Scandinavia, France, northern Italy, Rumania, western Russia, ? Spain. Introduced into North America. Material examined Living from many places in Britain. Spirit material from Britain, France and Italy. 130 BRITISH SLUGS Arion hortensis Ferussac, 1819 Arion hortensis Ferussac, 1819: 65, pi. 2, figs. 4-6. Presumably France. Arion elongatus Collinge, 1894 : 66, pi. 5A, figs. 1-4. Arion coeruleus Collinge, 18976 : 444. First British record By Gray (1821 : 239). Diagnostic features Differs from A . fasciatus in the following : body semicircular in transverse section when contracted, no enlarged mid-dorsal tubercles, reproductive orifice further back, mantle band surrounds respiratory orifice, sole yellow or orange ; spermatheca globular, free oviduct longer and proximally more slender, spermatophore hooked posteriorly and either discoidal or pointed anteriorly, brown patch on buccal bulb. External appearance (PL i, fig. 4) Length extended 2-5-3 cm - When contracted, the body is semicircular, not bell- shaped, in transverse section. The reproductive orifice is below the posterior lip of the respiratory orifice. The dark brown or black back is sprinkled with yellowish- brown dots, giving a pepper-and-salt appearance, the sole is pale yellow to bright orange, and the mucus is yellow. In alcohol, the back becomes bluish-grey and the sole loses its orange colour. The lateral bands are placed lower on the body than in A. fasciatus, and run along the edges of the mantle, surrounding the respiratory orifice instead of arching over it. External variation Var. grisea Moquin-Tandon (1855), P a ^ e g re Y an d bandless ; var. nigra Moquin- Tandon, bands nearly obliterated by general dark pigmentation ; var. rufescens Moquin-Tandon, reddish with black bands ; var. caerulea Collinge (1892^), bluish with dark bands, and yellow sides. Pallial organs As in A . ater. Retractor muscles As in A . fasciatus. Nervous system As in A . ater. Jaw (Text-fig. 7 G) About i mm across, doubtfully distinct from that of A. fasciatus. BRITISH SLUGS 131 Radula Similar to that of A. fasciatus, but the ectocone of the marginal teeth is less frequently present. Formula .11.24 x IO - Alimentary system (Text-fig. 4 A) Visceral mass less twisted than in A. ater. The buccal bulb has a brown shield- shaped mark on the upper surface an apparently good specific character not recorded in the literature. Reproductive system (Text-figs. 4 B and c) The very darkly pigmented ovotestis lies over the apex of the stomach, near the posterior end of the visceral mass, and is largely exposed on the surface. The middle dilated portion of the hermaphrodite duct is much folded, yellowish- white, and the terminal slender part shows the usual " claw " or seminal vesicle where it enters the elongated albumen gland. The sperm-oviduct is pale pinkish-brown, with a conspicuous yellow ribbon-like prostate. The free oviduct is unusually long and slender proximally, and dilated distally before entering the small white upper atrium. The vas deferens is short, and the epiphallus about the same length. The spermatheca duct is short and stout, with a bulbous base, and the sac is spherical. The lower atrium is yellow, with a soft glandular covering. As in A. fasciatus and A. intermedius, there is no ligula in the upper atrium. However, within the dilated distal end of the oviduct (Text-fig. 4 c), two conspicuous longitudinal folds correspond functionally with this structure during coitus (Quick, 1946). Below the termination of the spermatheca duct is a region of thickened mucous membrane in the atrium. In the slugs observed in Swansea (see below), the entering spermatophore adhered to this region. The epiphallus terminates in a short papillate papilla, which is very conspicuous when the atria are everted in coitus (Text-fig. 4 E). Both British and Dutch slugs show some variation in the folds of the oviduct representing the ligula, and in the relative length and thickness of the firm carti- laginous part ; but these differences could be due to differences in the degree of maturity of the animals, and to the extent of contraction during fixation in alcohol. Spermatophore (Text-figs. 4 F and M) Two types of spermatophore have been found in slugs from different localities. Since the spermatophores of Arion are seldom seen, it seems better to regard the slugs bearing the two types as varieties or local races, rather than as distinct species, until further anatomical or ecological differences are demonstrated. Slugs from Reading and from Zaandam in Holland (collected by Mr. D. Aten) have spermatophores in the form of simple chitinous tubes, 6-5 mm long, recurved at the slender posterior end and slightly shouldered near the anterior end, smooth and without denticulation (Text-fig. 4 F). Slugs from a garden in Eaton Crescent, Swansea (Glamorgan), which otherwise agree with those from Reading, have the very different spermatophores described 132 BRITISH SLUGS by Quick (1946). These are 5 mm long, shaped like cloves, with the anterior end forming a flattened disc with a crenulated edge, and with a ridge carrying about fifty-six forwardly pointing denticulations running the whole length of the convex side (Text-fig. 4 M). Mating (Text-fig. 4 D) Mating occurs in April and December, and probably in other months. After the usual " following" and licking of mucus, slugs from Swansea (spermatophores as in Text-fig. 4 M) curve in a semicircle head to head, and the atria are everted as a translucent mass with the lower atria stretched to a narrow yellow ring and the lower atria in contact. A long slender mobile club-shaped process is everted, which plays over the partner's body in the same manner as the sarcobelum of Agriolimax. However, subsequent dissection shows it to be the everted slender proximal part of the oviduct, functioning as a stimulatory organ, a modification which appears to be unique (Text-fig. 4 E). Coitus in Swansea slugs lasts for about an hour and a half. Gerhardt (1935) describes the mating of A. hortensis as lasting only twenty-five minutes, and the everted masses as small and featureless, in slugs whose spermatophores lacked serrations. This suggests that more than one species may be involved, though Mr. D. Aten (in litt.) states that the Zaandam slugs, having smooth spermatophores, evert their oviducts in coitus. Development The opaque yellow, sticky eggs measure 2-5 X 2 mm, and are laid in clusters of up to thirty. A second and third clutch may follow at three-weekly intervals. In the winter, the eggs take six or seven weeks to hatch. The newly hatched young are 4 or 5 mm long, and differ from those of A . fasciatus in having a darker dorsum and central mantle area. The usual body bands and lyriform mantle bands are present, and there is no dorsal line of pale conspicuous tubercles. The tentacles are violet-brown. Ecology This species is very common in gardens, allotments and fields, and burrows in the ground more than the other species of Avion. It can become a serious pest, by eating carrot, potato, etc. It is also frequent in woodland under logs and debris. Distribution (Map 7) Every county in the British Isles ; Scandinavia and western Europe to western Russia, (North Africa). Introduced into North America. Material examined Living and in spirit from many places in England and Wales, and from Zaandam, Holland. BRITISH SLUGS 133 Arion subfuscus (Draparnaud, 1805) Limax subfuscus Draparnaud, 1805 : 125, pi. 9, fig. 8. France. Arion fuscatus F6russac, 1819 : 65, pi. 2, fig. 7. First British record Possibly by Johnston (1838), as Arion subflavus (a nomen nudum). Up to about 1880, however, A. subfuscus was generally considered as a variety of A. ater. Diagnostic features Adults larger (6-7 cm) than A.fasciatus and A . hortensis, smaller than A . lusitanicus and A . ater ; differ from young of large species in smaller tubercles, narrower foot fringe, inability to contract to hemispherical shape. Internally differs from A . hortensis in relatively shorter proximal, slender, part of free oviduct, longer spermatheca duct. External appearance (PI. i, fig. 2) Length extended, from 5 to 7 cm. The colour is usually yellowish-brown, darker on the dorsum and head and on the body bands. On the mantle the right band completely surrounds the respiratory orifice. The foot fringe is yellowish and lineo- lated more or less distinctly with black, and the sole is pale yellow, sprinkled with whitish specks, and obscurely tripartite. The mucus, especially on the mantle, is yellow or orange. When contracted, A . subfuscus neither assumes the hemispherical shape nor exhibits the rolling reflex (p. 145) of A. ater. External variation This as usual depends on the relative development of black, red and yellow pigment, and the distinctness of the bands : var. rufofusca Draparnaud (1805), rufous with black bands ; var. bicolor Moquin-Tandon (1855), red with darker dorsum and no bands, foot fringe orange ; var . fuliginea Morelet (1845), dark brown, fringe yellowish; var. nigricans Pollonera (18876) , black ; var. cinereofusca Draparnaud (1805), ashy grey ; var. succinea Bouillet (1835), yellow or orange ; var. pegorarii Lessona & Pollonera (1882), quadrifasciate ; var. alba Esmark (1883), whitish. In most cases, the yellow pigment is largely contained in the mucus, and when this is removed the animal becomes brown or grey, as do alcohol specimens. Pallial organs, retractor muscles, nervous system As in A . ater. Jaw Brown, arcuate, about 1-5 mm across, with from ten to sixteen transverse ribs. 134 BRITISH SLUGS Radula (Text-fig. 3 H) About 4-5 X 2 mm. About ten more teeth on each side of each row than in A. hortensis. Formula .15.30 x 140. Alimentary system Similar to that of the other species of Avion, except that the visceral mass is rather more twisted than in A . intermedium and A . fasciatus, and less so than in A. ater. Reproductive system (Text-figs. 6 D and F) From the pale brown ovotestis, the greatly convoluted white hermaphrodite duct enters the albumen gland, with the usual "claw" or seminal vesicle. The common duct is voluminous and folded, the male or prostatic portion yellow and the female part pinkish-grey. The first part of the free oviduct is short and slender, and the longer distal part dilated. This part contains two prominent longitudinal folds of mucous membrane forming the ligula as in A. hortensis. The spermatheca duct is bulbous at its origin, then longer and more slender than in A. hortensis, and ends in a larger spherical spermatheca sac. The vas deferens and epiphallus are very similar to those of A. hortensis, but the terminal annular thickening is more pronounced. The upper atrium is small, the yellow glandular lower atrium of medium size with a wrinkled mucous membrane. Spermatophore (Text-fig. 6 G) Very different from that of A . hortensis : about 20 mm Ipng, slender anteriorly, then dilating, to taper again to an almost filamentous tail. The convex edge is finely denticulate. Mating Mating occurs in March, April and May, and perhaps other months. One of a pair follows the other for about half an hour, " licking " the caudal mucus. Then they curve around in a semicircle with the right side of the necks in contact. In about fifteen minutes the lower atria are everted as small yellow eminences. These rapidly enlarge as eversion is completed, and the combined atria form an ovoid bluish-white mass between the two animals, the junction of the two atria being barely visible as a fine line. After three-quarters of an hour the mass shrinks and the animals separate, and the spermatophores are momentarily visible, partly protruding from the spermatheca ducts as the remainder of the shrinking mass is withdrawn. For further details see Quick (1946). The slender portion of the free oviduct is not everted as a stimulator as it is in A. hortensis, where this portion of the oviduct is longer. The combined mass resembles that of A. ater, since the differing situation of the ligula does not affect the aspect. BRITISH SLUGS 135 Development The opaque leathery eggs, measuring from 2-25 X 2-25 to 3-25 X 2-25 mm, vary considerably in size and shape, even in the same cluster. Eggs are laid three weeks after mating. They are laid in masses of up to fifty, and tend to adhere together, several eggs sometimes being enclosed in a common tubular coat of mucus. They are white at first and become yellow or brown. Hatching occurs in from three to five weeks. The newly hatched young are about 6 mm long, grey, with darker dorsal region, and with lateral body and mantle bands. The tentacles are violet. Newly hatched A . ater are orange, straw colour or pale greenish-grey, and the tentacles dark grey rather than violet ; but the most notable difference is that the anterior ends of the mantle bands are closer together than in A. subfuscus (PI. i, figs. 9 and ii). Ecology A. subfuscus lives under logs and litter in deciduous and coniferous woods, and is one of the few species found in pine woods. It is also frequent in damp hedgerows, waste ground and some gardens. It feeds on the algal and fungal film on logs, and also on the larger fungi. In captivity it readily eats carrot, lettuce, oatmeal and dead slugs (Map 8). Distribution The whole of Great Britain and Ireland, except possibly parts of East Anglia and eastern Scotland. Most of Europe, north of a line through Orenburg and Odessa (Russia), northern Italy, Barcelona and Oporto. Introduced into North America. Material examined Living from many places in Britain. Spirit material from Britain, France, Luxem- bourg, Germany, Italy and Madeira. Arion lusitanicus Mabille, 1868 Avion lusitanicus Mabille, 1868 : 134. La Sierra d'Arriba, near Lisbon. Notes The name lusitanicus is applied to this species with some hesitation, since Mabille (1868) described the tubercles as " peu allonges ". However, Pollonera's description and figures (1889 : 626, pi. 9) leave little doubt about the identification. A. nobrei Pollonera (1889) is similar to A. lusitanicus but smaller, with smaller tubercles, completely black, and with the distal genital ducts even more strongly pigmented. A specimen from Portugal in the B.M. (N.H.) collections, identified as A. nobrei, agrees with this description. Cain & Williamson (1958) discuss specimens of Arion collected at Nuneaton, Warwickshire, which perhaps suggest that the status of A. lusitanicus in Britain may require further investigation. ZOOL. 6, 3. 10 136 BRITISH SLUGS First British record By Collinge (1893). Though again recorded by Collinge (1897^), the species was overlooked by Taylor (1902-07) and by other authors on the British fauna (e.g. Roebuck, 1921 ; Kennard & Woodward, 1926 ; Ellis, 1926 and 1951 ; Quick, 1949), until its rediscovery in Durham by Mrs. Vincent (Quick, 1952). Diagnostic features Large, up to 10 cm in length, with coarse tubercles like A . ater ; from which it is internally distinguished by the dilated distal part of the free oviduct, which lodges the ligula. Distinguished from A . subfuscus by coarse tubercles and a longer sperma- theca duct. External appearance The length when crawling is from 7 to 10 cm. The colour varies as in A. ater, and may be dark grey, reddish-yellow, brown or dull greenish-grey, with darker head and tentacles. The transversely lineated foot fringe is sometimes darker and sometimes lighter than the body, and the sole is generally pale, but dark in some Nuneaton specimens. The respiratory orifice does not dilate to the same extreme degree as in A. ater. In the younger stages there is on each side a dark lateral band (Text-fig. 5 D). On the mantle the bands form a lyre-shaped figure, the right band surrounding the pulmonary orifice on all sides, and anteriorly they are a little closer together than in A. subfuscus of similar size. Sometimes the bands are retained up to maturity. External variation Var. rufescens Collinge (1893), body dark red, sides of body bandless ; var. nigres- cens Collinge, black or plumbaceous ; var. olivaceus Collinge, various shades of olive-green ; var. flavogriseus Collinge, yellowish-grey, foot fringe lighter than the body. Pallial organs, retractor muscles, nervous system As in A . ater. Jaw Intermediate in size (3 mm across) between those of A . subfuscus and A . ater ; arcuate, crossed by from nine to thirteen rather broad even ribs ; pale brown or yellow in light examples, nearly black in dark ones. i Radula (Text-fig. 3 G) Intermediate in size between those of A. subfuscus and A. ater. The formula varies (in adult specimens from Durham) between C.I7.35 and C.I9.45. Alimentary system As in A. ater. BRITISH SLUGS 137 Reproductive system (Text-figs. 5 A and E) The very dark ovotestis lies a little further forward than in A. ater. The dilated central part of the hermaphrodite duct is folded in a zig-zag manner, and then narrows to form the usual " claw " at the albumen gland, which varies in size with the maturity of the animal. A yellowish-pink fold or ruffle accompanies the first part of the sperm-oviduct. The free oviduct is slender at first, and then becomes greatly enlarged, to narrow again slightly where it enters the upper atrium. This dilated part, in both the Durham and the Nuneaton examples, is deeply pigmented in a characteristic way (Text-fig. 5 A). The spermatheca duct is rather longer than in A. subfuscus, somewhat dilated at the origin. It also sometimes shows a slight dilatation, with stronger mucous folds, between the origin and the spherical sperma- theca sac. The vas deferens is about as long as the epiphallus, and the latter shows a conspicuous annular ring at its termination. The distal half of the epiphallus is deeply pigmented like the oviduct, the pigment usually tailing off in two tongues. In some individuals this pigmentation is less pronounced. The upper atrium is small, the lower pear-shaped with thick glandular-looking walls, sometimes yellow, and sometimes sprinkled with minute spherical white granules. The upper atrium contains no ligula, but the distal dilated part of the oviduct contains two conspicuous folds which apparently represent this structure, and function similarly (Text-fig. 5 E). In A. ater the ligula occupies the large upper atrium in both young and adult animals. Examination of the genital ducts of imma- ture animals of A. subfuscus and A. lusitanicus (Text-figs. 6 c and D) shows that the ligula lies in the oviduct, as in the adult, and does not " migrate " upwards or downwards. It is noteworthy that in young A. ater the oviducal branch of the genital retractor is inserted as in A. hortensis, A. subfuscus and A. lusitanicus ; while in adult A . ater (Text-fig. 6 K) the distal part of the oviduct becomes relatively much shorter, bringing the retractor close to the atrium. It could be argued that the lateral atrial pouch of A. ater is homologous with the dilated terminal portion of the oviduct of the other species, or alternatively that this portion of the oviduct is a diverticulum of the atrium. Seeing that this portion is so well devleoped in the early stages of growth, the second alternative is unlikely. There is in Geomalacus maculosus a long backward extension of the atrium ; but this carries back the epiphallus and the spermatheca, and does not affect the oviduct. Spermatophore (Text-figs. 6 H and i) A chitinous tube 20 mm long, armed with a spiral row of serrations, most prominent towards the anterior end ; and two low smooth ridges at the posterior end, which die out anteriorly. Mating (Text-fig. 6 E) The figure shows the appearance of the reproductive organs just before they are withdrawn after mating, with the partner's spermatophore protruding from the spermatheca duct. Observations on the mating of A. lusitanicus are too few and incomplete to permit of comparison with the other species. 138 BRITISH SLUGS Development The white eggs measure about 3 X 3-25 mm, but vary considerably in size and shape, both within and between clusters, and become yellower with age. In the laboratory, they have been laid in January and in March ; and in September by laboratory-bred animals only five or six months old. The clusters contain up to about fifty eggs, bound together by mucus, which hatch in about three weeks in the laboratory. Ecology In captivity, like most other slugs, A. lusitanicus is readily reared on lettuce, cabbage, oatmeal, etc.; but little is known of its habits in the wild. Distribution (Map 9) In the British Isles, known only from Berehaven, Bantry Bay, Ireland (Collinge, 1893) ; the garden of No. 24, North Bailey Road, Durham and the river bank adjoining (Quick, 1952) ; and the site of bombed cottages at Attleborough, Nuneaton, Warwickshire (Cain & Williamson, 1958). On the Continent, known from near Lisbon, Portugal ; Mentone and Pyrenees Orientales, France ; and Basle, Switzer- land. The species may be indigenous to Ireland and but recently introduced to Durham and Nuneaton, though it seems firmly established there. Probably it is more widely distributed in Britain, but has not been distinguished from A. ater. Material examined Living specimens from Durham and Nuneaton. Spirit specimens from Basle, Switzerland, and the Pyrenees Orientales. An immature specimen from Bantry Bay (Collinge Collection, Museum of Zoology, Cambridge) probably belongs to this species. Arion ater ater (Linnaeus, 1758) Limax ater Linnaeus, 1758 : 652. Sweden. Arion empiricorum Ferussac, 1819 : 60, pis. 1-3 (in part). First British record By Lister (1674). Diagnostic features Large size, coarse tubercles, wide foot fringe, large respiratory orifice ; short free oviduct surrounded by genital retractor muscle, ligula in upper atrium. Adults not certainly distinguishable in outward appearance from A. a. rufus, but usually black or dark brown with a grey sole, and rarely with a brightly-coloured foot fringe. Lateral pouch and ligula much smaller than in A . a. rufus, and vas deferens usually shorter. Upper atrium shorter than lower atrium. BRITISH SLUGS 139 External appearance (PL i, fig. 3) Length when fully extended generally about 14 cm, but may be even longer. The body tubercles are long and prominent, and darker than the furrows except in the young. The mantle is shagreened. There is no trace of a dorsal keel at any stage. The caudal mucus pore is conspicuous. The sole is more or less divided into median and lateral areas, the lateral areas usually the darker. The wide foot fringe is transversely lineolated, and in dark specimens the lines extend into the lateral areas of the sole. In pale individuals (except true albinos, which are rare) the head and tentacles are much darker than the rest of the body. The respiratory orifice is absolutely and relatively larger than in the previously described species. The very sticky body mucus may be colourless, or yellowish in red and yellow forms, that from the caudal pore always colourless. At rest, A. ater, like A. lusitanicus, but unlike A. subfuscus and the smaller species, can contract to a hemispherical shape, or even a larger portion of a sphere. External variation Since A. a. rufus has been generally considered as only a variety of A. ater, it is impossible at present to tell which of the very numerous named varieties belong properly to the latter. Probably most of the brighter varieties belong to A . a. rufus, and of the duller ones to A . a. ater. Extensive lists of varieties are given by Taylor (1902-07), Kennard & Woodward (1926) and Hesse (1926), of which the following is a short selection : var. ater Linnaeus (1758), black with grey sole ; var. aterrima Taylor (1905), completely black ; var. alba Linnaeus (1758), white ; var. albida Roebuck (1883), white with yellow fringe ; var. castanea Dumont & Mortillet (1856), brown with paler fringe ; var. albolateralis Roebuck (1883) (? = var. media Jensen, 1873), dorsum black, sides white ; var. reticulata Roebuck (1885), tubercles pale, furrows pigmented (i.e. the juvenile colouring retained by the adult) ; var. hiberna Mabille (1868), rusty purple with paler sides ; var. plumbea Roebuck (18840), leaden grey with yellow fringe ; var. succinea Miiller (1774), yellow with red or orange fringe ; var. coccinea Gistel (1848), vermilion ; var. marginella Schrank (1803), black with red fringe. Pallial organs Large vessels occupy the roof and floor of the pulmonary chamber. The kidney is curved into a rounded quadrangular form, surrounding the pericardium, and the ventricle lies posteriorly to the auricle. The anterior branch of the aorta curves round the anterior loop of the intestine, then runs forwards supplying branches to the anterior part of the body, passing between the visceral and pedal ganglia. The posterior aorta runs backwards and supplies the alimentary and genital systems, and passes between the visceral and pedal ganglia. The ureter and rectum terminate at the posterior lip of the respiratory orifice. 140 BRITISH SLUGS FIG. 4. Avion. A. A . hortensis internal organs in situ ( x 2) showing liver, exposed part of stomach, part of ovotestis, albumen gland, intestine,, aorta, crop, left salivary gland, nerve ring, characteristic pigmented area on buccal bulb. B. A. hortensis distal genital ducts ( x 4) with position of the partner's spermatophore indicated on upper atrium. G. A. hortensis distal portion of genital ducts (x 5-3) opened to show ligula in oviduct, glandular area below spermatheca, and papillate mucous membrane of epiphallus. D. A. hortensis mating pair (x 1-3) with slender parts of oviducts everted and functioning as stimulators. E. A. hortensis individual BRITISH SLUGS 141 B FIG. 5. Arion. A. A. lusitanicus reproductive organs (x 3). B. A. subfuscus distal genital ducts (x 3). G. A. ater ater distal genital ducts (x 3). D. A. lusitanicus- young individual ( x 10). Compare mantle bands with those of young A. subfuscus and A. ater (pi. i, figs. 9 & n). E. A. lusitanicus oviduct (x 5-3) opened to show position of ligula. immediately after coitus ( x 2) with everted genitalia, showing atria, everted oviduct, papillate epiphallus, partner's spermatophore adhering to base of spermatheca duct. F. A. hortensis spermatophore ( x 6-7). G. A.fasciatus spermatophore ( x 4), adapted from Lohmander. H. A. intermediusdistal genital ducts (x 4-7). I. A. fasciatus distal genital ducts (x 4). K. A. hortensis pallial region in ventral view, showing kidney, pericardium, auricle, ventricle, rectum and respiratory orifice. Muscles cut short. L. A . hortensis cephalic and buccal retractor muscles, ventral view. M. A. ? hortensis spermatophore (x 6-7) of a slug from a garden in Eaton Crescent, Swansea, Glamorgan. I 4 2 BRITISH SLUGS A. FIG. 6. BRITISH SLUGS 143 Retractor muscles The right and left cephalic retractors are widely separated at their origin, from the diaphragm behind the kidney. The pharyngeal retractor arises from the median dorsal body-wall. The genital retractor arises from the diaphragm immediately in front of the origin of the left cephalic retractor, ensheathes the short free oviduct, and sends a branch to the spermatheca duct. Nervous system (see Text-fig. 7 M) Highly concentrated in the adult, the commissures being short and the visceral ganglia closely united. Jaw Strong, arcuate, 3-5 mm across, crossed by a varying number of ribs (up to eighteen), variable in breadth, which denticulate the upper and lower edges. Colour yellow in pale specimens, nearly black in dark ones, the attached margin being darker than the free cutting edge. Radula (Text-fig. 3 F) 7-8 mm long by 3 mm wide. The number of lateral teeth varies from 20 to 25, and of marginals from 40 to 46 on each side, but the average formula is C.22 . 44 X 130. Alimentary system The short oesophagus leads through a large wide crop to the stomach, which forms the posterior part of the alimentary canal. The salivary glands on each side of the crop are large, flattened and much lobulated. The intestine, with one anterior loop, is twisted through about one and a half turns more than in the young of the smaller species. Reproductive system (Text-figs. 5 c and 6 L) The proximal parts resemble those of other species of A rion (see Text-fig. 5 A). FIG. 6. Avion. A. A. ater ater distal genital ducts (x 10) of a young individual, showing long atrium and position of ligula. B. A . ater ater genital organs ( x 3-3) of a nearly full-grown but immature individual, showing relative sizes of ovotestis, albumen gland, common duct and distal portions of ducts. C. A, lusitanicus genital organs ( x 6-7) of an immature individual, showing relative sizes of organs and position of ligula. D. A. subfuscus distal genital ducts (x 6-7) of a young individual, showing ligula in lower oviduct. E. A. lusitanicus slug after coitus, with atrium everted. The ligula, with the oviducal orifice in its lower anterior part, surrounds the orifices of epiphallus and spermatheca. The partner's spermatophore projects from the spermatheca duct. F. A . subfuscus distal genital ducts ( x 3-3). G. A. subfuscus spermatophore (x 3-3), and portions of spermatophore (x 16-7) from near anterior end, centre and posterior end. H. A. lusitanicus spermato- phore (x 3-5). I. A. lusitanicus portion of posterior end of spermatophore (x n). K. A. ater rufus distal genital ducts (x 3-3). L. A, ater ater distal genital ducts (x 3-3). 144 BRITISH SLUGS The short free oviduct, ensheathed by the genital retractor, enters the upper atrium apically. It is not dilated, and does not contain the two conspicuous folds of A. hortensis, A. subfuscus and A. lusitanicus. However, the upper atrium contains a somewhat similar folded mass, arising from the walls of the atrium and the termina- tion of the oviduct, known as the ligula. This is much smaller than in A. a. rufus, where it is contained in a lateral enlargement of the atrium. The spermatheca duct is relatively longer than in A. intermedium, A. fasciatus, A. hortensis and A. subfuscus, is not bulbous at the base, and receives a branch from the genital retractor. The vas deferens is usually less than one and a half times as long as the epiphallus, which terminates in a thickened ring at the atrium. Spermatophore (see Text-fig. 7 L) This rapidly disintegrates in the spermatheca, so that it can only be obtained intact soon after coitus. Two specimens were obtained in this way, which are 17 and 18 mm long. Each is somewhat dilated near the proximal end, and tapers in each direction. A finely serrated ridge runs nearly the whole length, the serrations pointing forwards. The epiphallus measures only n mm long in alcohol specimens ; so, unless it is much longer in life, part of the spermatophore must be formed during coitus. Mating Occurs in June and July, at any rate in Wales, after dark. After about half an hour of following and licking, the animals curve in a semicircle, neck to neck, and evert their atria to form a bluish- white mass. They separate after about half or three-quarters of an hour, when the spermatophores can be seen protruding from the spermatheca ducts, while the shrinking ligulae and atria are quickly withdrawn. Adams (1910), Gerhardt (1940) and Quick (1947) give fuller details. Development The eggs are variable in size, usually about 5x4 mm. They are opaque and leathery, pearly white when laid but turning brown. They are laid in clusters in the soil, two or three weeks after coitus. The first cluster may contain 150 eggs or more, and one or two subsequent clusters fewer than this. The eggs hatch in from four to six weeks. The young, 10 mm long, are pale straw colour or sometimes greenish, with darker head or tentacles. They are usually bandless, and if faint bands appear after a few days they quickly fade. Yellow pigment is concentrated in the tubercles. In the young of dark varieties the black pigmentation begins in the dorsal furrows and spreads downwards, producing a reticulate pattern which very occasionally persists in the adult ; though as a rule the pigment of pale adults is concentrated in the tubercles, leaving the furrows paler. The sole is the last part to pigment. Like the other species of Arion, A. ater shows in the later stages of embryonic development a cephalic and a caudal vesicle, which pulsate alternately. The tubular caudal vesicle persists to within a few days of hatching (Text-fig. 18 F). The jaw at hatching does not show any sign of bilateral origin like that of Agriolimax. BRITISH SLUGS 145 Behaviour Adults of A . ater exhibit a peculiar rolling reflex, if touched when fully (or almost fully) contracted. The body sways from side to side with a simultaneous screwing motion, with a periodicity of about two seconds. The movement cannot always be elicited, but may be quite violent. As far as is known, no other slugs show this reflex. It is tempting to explain it in terms of alternate contraction of the cephalic retractors, which have separated origins in Avion. However, the reflex is only doubtfully elicited from A . lusitanicus, and repeated attempts to elicit it from other British species of A rion have always failed. Ecology A. a. ater is practically omnivorous, eating decayed vegetable matter, dead mice and rabbits, and animal faeces. In the lowlands, it is common in grass fields, woods, hedgerows and waste ground, and on grassy downs and sea cliffs. It is often the only mollusc present on upland peaty moors and sphagnum bogs, where it is smaller than in the lowlands, and either black (often the var. aterrima) or very dark brown. Distribution (Map 10) The whole of the British Isles, northern and central Europe, Iceland and ? Russia. Probably some of the records from Spain, Portugal, Italy and the Mediterranean coasts and Atlantic isles refer to the related forms (A. a. rufus, A. lusitanicus and possibly others). Introduced into North America. Material examined Living specimens of A. a. ater, as distinct from A. a. rufus, have been identified from the following localities in the British Isles : Exeter, Littlehampton and Collompton in Devon ; Netley Heath, Surrey ; Howe Wood near Littlebury, Essex ; Reading, Berkshire ; Thorpe St. Andrews, Brundall Marshes and Wheatfen Broad in Norfolk ; Little Shelf ord, Homingsea and Durnford Fen in Cambridgeshire ; Cefn Bryn, Gower and Llangennith in Glamorgan ; Capel-y-ffin, Brecon ; Skokholm Island, Pembroke ; Abersoch, Caernarvon ; Bromborough, Cheshire ; Pentland Hills, Edinburgh ; Courtmacsherry, Cork. Arion ater rufus (Linnaeus, 1758) Limax rufus Linnaeus, 1758 : 65. ? Sweden. Arion empiricorum Ferussac, 1819 : 60 (in part). Arion sulcatus Morelet, 1845 : 28. Notes See Quick (1947). 0kland (1923) believes it to have been introduced to Sweden in historical times. Ferussac's empiricorum includes A. a. ater, A. a. rufus, and perhaps also A . lusitanicus and other species. Moquin-Tandon's description (1855-56, 2 : 10) of empiricorum, and his fig. i on pi. i, could serve for either A. a. ater or 146 BRITISH SLUGS A. a. rufus ; but his fig. 12 represents the genitalia of A. subfuscus, A. lusitanicus or a similar species. Cain & Williamson (1958) have studied pigmentation and genital anatomy in populations referable to A. a. ater and A. a. rufus. Their evidence suggests that only the former is indigenous to the British Isles, where A. a. rufus is an introduced alien associated with cultivated land. In Europe the natural ranges of the two forms seem not to overlap widely, since A. a. ater apparently occupies Scandinavia and A. a. rufus much of the rest of the continent, but it is not yet known how the two forms interact along their natural boundary. However, two populations near Oxford show the intermediacy of character combined with high variability which is to be expected from hybrids. This suggests that the two forms should be considered as geographical races. First British record By Gerhardt (1940), on specimens from Swansea, Glamorgan. The subspecies had been considered by previous British authors as merely a variety of A . a, ater. Diagnostic features Adult not certainly distinguishable from A. a. ater on external appearance, but usually yellow to reddish-brown, with a creamy yellow sole and bright orange or red foot fringe. Internally, upper atrium longer than lower, much larger ligula in lateral pouch, vas deferens usually longer than in A. a. ater. Young have mantle bands, rarely found in A . a. ater. External appearance, external variation, pallial organs, retractor muscles, nervous system, jaw, radula, alimentary system As in A . a. ater. Reproductive system (Text-fig. 6 K) As in A. a. ater, except for the distal genital ducts. The upper atrium is larger, and exhibits a lateral bulging that lodges a much larger ligula. The lower atrium is shorter than the upper. The vas deferens is relatively longer, usually more than one and a half times as long as the epiphallus and sometimes nearly twice as long. The oviduct, instead of entering the apex of the atrium formed by the lateral bulge, enters between this and the entrance of the epiphallus. Sometimes (as in specimens from Limeslade Bay, Swansea) the lateral bulge is narrower and prolonged further backwards. Occasional specimens of A . a. ater show a suggestion of a lateral bulge, but the size of the ligula, and the relative sizes of upper and lower atria and the vas deferens and epiphallus, serve to identify it. Spermatophore (Text-fig. 7 L) Probably longer than that of A. a. ater (about 25 mm long), and even more disproportionate to the size of the epiphallus in spirit material. This may be associ- ated with the longer period of coitus, if the spermatophore is formed during coitus. BRITISH SLUGS 147 Mating A . a. rufus attains maturity later in the year than A . a. ater, and mates in September and October, at any rate in South Wales. Mating resembles that of A. a. ater, except that coitus lasts for two hours. For further details, see Gerhardt (1940) and Quick (1947). Development The eggs are like those of A. a. ater. The newly-hatched young are pale yellow or orange, and differ from those of A. a. ater in having (at birth or after a few days) dark mantle and body bands, which persist for a few months. In individuals which will be dark when adult, the course of pigmentation is as in A. a. ater. The young can be distinguished from those of A . subfuscus by their coarser tubercles, and by the closer approximation anteriorly of their mantle bands (PI. i, figs. 9 and n). Behaviour A. a. rufus shows the same rolling reflex as A. a. ater. Ecology In Britain, A. a. rufus is more likely to be found in gardens and parks (though sometimes at a considerable distance from dwellings) than in wilder places. It never seems to occur in peaty moors or sphagnum bogs ; though this may reflect its introduction by man, and consequent restriction to cultivated land, rather than dependence on calcium in the soil. Distribution (Map n) Not fully known, since A. a. rufus has only recently been distinguished sub- specifically from A. ater. Authentically known in Britain from Glamorgan, Somerset, Berkshire, Oxford, Essex, Cambridge and Leicestershire. It seems to be the native form in Europe south of Denmark, though the natural ranges of the two subspecies are insufficiently known. Material examined Living specimens from Monkton Coombe near Bath, Somerset ; Chingford, Essex ; Basildon, Berks ; Coombe End, Oxford ; Lords Meadow near Cambridge ; and the Swansea district, Glamorgan. Spirit material from Loughborough, Leicester- shire ; Borgholm, Oland Island, Sweden ; Froedensborg, NE. Zealand, Denmark ; Canton Solothurn, Switzerland ; and from France, Corsica and Portugal. Superfamily ZONITACEA Family LIMACIDAE Slugs with a small calcareous shell almost always completely enclosed by the mantle. The dorsum is keeled posteriorly or right up to the mantle. There is no I 4 8 BRITISH SLUGS caudal mucus gland. The sole is tripartite and a supra-pedal groove is present (aulocopod). The jaw is smooth, usually with a median projection (oxygnath), and the lateral radular teeth are tri- or bicuspid, the marginal teeth aculeate. The spermatheca duct is short, generally arising from the atrium. FIG. 7. Limacidae. External appearance. A. Milax 'keel extending forward to mantle ; mantle granulate with horseshoe furrow ; respiratory orifice behind centre of mantle ; genital orifice between upper tentacle and respiratory orifice. B. Agriolimax keel extending only a little way forward ; tail truncate ; mantle con- centrically ridged with nucleus over respiratory orifice, behind centre of mantle ; genital orifice close behind upper tentacle. C. Limax keel extending a little way forward ; tail not truncate ; mantle concentrically ridged with nucleus in median dorsal line ; respiratory orifice behind centre of mantle ; genital orifice close behind upper tentacle. D. Milax chevroned foot-sole, stimulator protruded. E. Lehmannia marginata jaw. F. Agriolimax reticulatus jaw of very young individual, showing bilateral origin. G. Avion hortensis jaw, for comparison. H. Limax maximus shell. Avion ater rufus. I. Pair in copula. K. Portion of posterior end of sperma- tophore (x 24). L. Spermatophore (x 1-7). M. Ganglia and commissures of nerve ring (buccal ganglia omitted). BRITISH SLUGS 149 Subfamily PARMACELLINAE The shell is paucispiral, not completely enclosed by the mantle (in Parmacella], or flat and enclosed (in Milax). The mantle is granular, bearing a horseshoe-shaped groove, and the respiratory orifice lies behind the middle of the right margin. The genital orifice is between the respiratory orifice and right upper tentacles. The dorsum is keeled up to the mantle. The central and lateral radular teeth are tricuspid. The intestine has one forwardly directed loop, and the rectum lacks diverticula. An epiphallus secretes a spermatophore, and atrial glands are present. Milax is the only British genus. Watson (1930) considers that Parmacella and Milax arose from a common stock ; Parmacella, with its spiral shell, embedded pedal gland, and ocular retractor crossing the penis, being the more primitive. Genus MILAX Gray, 1855 Type species Limax gagates Draparnaud, 1801. The shell, completely enclosed by the mantle, has a median non-spiral nucleus near the posterior margin. The pedal mucus gland lies free in the body cavity, and not embedded in the foot below it, as in the other genera. The right ocular retractor lies to the left of both penis and vagina. In M. sowerbyi and M. gagates an atrial stimulator is present. The median area of the sole is crossed by A-shaped grooves. Milax is without lateral body bands at all stages of growth. Watson (1930) considers that, of the British species of Milax, M. gagates (with more lateral and fewer marginal radular teeth, and a wider distribution) is the most primitive, and M. budapestensis (with fewer lateral and more marginal teeth, and no stimulator) the most specialized. A common continental species, M. marginatus (Draparnaud, 1801) (= M. msticus Millet, 1843), which it seems possible may yet be discovered in Britain, is a yellowish slug with small black spots on the body and mantle and a dark band on each side of the latter, and resembles M. budapestensis in lacking the atrial stimulator. Milax gagates (Draparnaud, 1801) Limax gagates Draparnaud, 1801 : 100. Presumably near Montpellier, France. Amalia parryi Collinge, 1 89513 : 7. Amalia babori Collinge, iSgjc : 294. Notes Two species are currently confused under the name gagates. The common British species is distinguished by its smooth genital stimulator, and less certainly by its smoother appearance (with flattened tubercles) and paler sole. The second species is primarily Mediterranean in distribution, though there is one hitherto unpublished 150 BRITISH SLUGS British record (p. 156). It has two rows of large papillae on the stimulator, and also more prominent tubercles and a darker sole. It is not easy to determine the proper application of the name gagates. There are no slugs among Draparnaud's types at the Naturhistorisches Museum, Vienna (Locard, 1895 : 154; confirmed by Dr. O. E. Paget in litt.}. Both species occur at Marseilles, only eighty miles from Draparnaud's home at Montpellier. Drapar- naud's description does not deal with the internal anatomy. However, he describes gagates as " tres-lisse, tres-luisante " (1801) and " nitidus ; corpore striato subrugoso" (1805). This strongly suggests the north-western species, since he applies " sub- rugosus " to other slugs as smooth as this, and uses " rugosus " only for those as rough as the Mediterranean species. Few early authors described or figured the condition of the stimulator, from which alone their application of the name gagates could be determined. Lessona & Pollonera (1884 : 105, pi. 2, fig. 2) use this name for Italian slugs with papillate stimulators. Germain (1930, fig. 52 A), in treating the French fauna, gives a figure of a papillate stimulator which is apparently copied from Lessona & Pollonera, and does not mention the locality. Taylor (1902-07, fig. 159) shows the smooth stimulator of a British slug. The name parryi Collinge is a synonym of gagates, as the latter is here interpreted, since one of the type specimens from Santa Cruz, Teneriffe, shows on dissection a penis, epiphallus and stimulator like the British species. So does one of a series from Haleakala Maui, Hawaii, apparently the types of babori Collinge. The following names, among others, cannot at present be allocated to either of the species with any certainty : Limax maurus Quoy & Gaimard, 1824 : 4 2 7- Port Jackson, Sydney, Australia. Parmacella nigricans Schultz, 1836, 1 : 125. Palermo, Sicily. Amalia mediterranea Cockerell, 1891 : 331. Algeria. Amalia mediterranea forma similis Cockerell, 1891 : 332. Catania, Sicily. However, specimens that I received from Catania in 1947 proved to be the species with a papillate stimulator. First British record By Thompson (1840 : 205). Diagnostic features Uniform grey or black colour, with darker keel, respiratory orifice without pale margin, mucus colourless ; smooth genital stimulator, rather thick spermatophore, not hooked at the thick end and with spines distributed more evenly than in M. sowerbyi. Differs from M. cf . insularis in smooth, not papillate, stimulator ; and less certainly in smoother skin and paler sole. External appearance (PL i, fig. 12) Length extended about 5 cm. When contracted, the slug is more compressed laterally than M. sowerbyi. The colour is usually grey or black, getting lighter BRITISH SLUGS 151 towards the foot. The tubercles are flattened, making the slug rather smooth, and the intervening grooves are unpigmented. The keel is usually dark, and is abruptly angulated at the posterior end. The central area of the granulate mantle is delimited by a horseshoe-shaped groove, open posteriorly, with the arms angulated at the level of the respiratory orifice. The respiratory orifice is small and without a pale border, and is situated a little behind the centre of the right margin of the mantle. The sole is pale, distinctly tripartite, with the median zone crossed by A-shaped grooves (Text-fig. 7 D). Only this central area shows the locomotory waves. The mucus is sticky and colourless. External variation Var. plumbea Moquin-Tandon (1855), paler or darker grey ; var. rava Williams (1888), drab coloured, slightly fuscous, mantle paler; var. bicolor Taylor (1904), deep red on the sides ; var. benoiti Lessona & Pollonera (1882), black, keel whitish. Other varieties with a Mediterranean distribution mentioned by Taylor (1902-07 : 143-146) probably refer to M. cf. insularis. Shell (Text-fig. 8 K) Measures about 4 X 2-5 mm. White and shining. Somewhat saddle-shaped above, and flattened below. The nucleus is near the posterior margin, in the middle line, and the lines of growth are distinct. Viewed laterally, the shell is concave from back to front, with the nucleus at the highest point, and shows a groove separat- ing the upper part of the shell from the lower. Pallial organs The kidney has an extension to the right toward the rectum, and the auricle and ventricle lie anterior to the left half of the kidney. The rectum crosses over the common origin of the tentacular and pharyngeal muscles to the posterior lip of the respiratory orifice. Pedal gland Lies free in the body cavity, occupying half the body length. Its duct projects posteriorly from the glandular tissue for about 0-25 mm. Retractor muscles The pharyngeal and tentacular retractors are not separated, as in Arion, but have a common origin from the posterior edge of the diaphragm in the middle line. At the posterior border of the kidney this divides into the pharyngeal retractor, which is bifurcate for about one-third of its length, and the tentacular retractor. The latter soon divides into right and left branches, and these divide again much further forward into branches to the upper and lower tentacles. The penial retractor arises further forward, on the left side of the diaphragm. ZOOL. 6, 3. II 152 BRITISH SLUGS FIG. 8. Milax. Distal genital ducts ( X 4). A. M. gagates atrium opened to show stimulator. B. M. sowerbyi atrium opened to show stimulator. C. M. buda- pestensis. D. M. cf. insularis. Spermatophore (x 6-7), spines (x 33). E. M. gagates one spine. F. M. sowerbyi two spines. G. M. budapestensis three spines. H. M. cf. insularis papillate stimulator. Shell (x 2-7), dorsal and lateral views. I. M. sowerbyi. K. M. gagates. L. M. budapestensis. BRITISH SLUGS 153 Nervous system The cerebral ganglia are connected by a short broad commissure. The cerebro- pleural and cerebro-pedal commissures are short, so that the nerve ring closely surrounds the oesophagus. The pleural, visceral and abdominal ganglia are closely approximated, as are also the visceral and pedal ganglia. The abdominal ganglion is almost completely fused with the left visceral. There is little if any variation in the nervous system between the species of Milax. Jaw Brown, oxygnathous, measuring 1-7 mm across. Radula (Text-fig. 9 A) The central tooth has a mesocone and well-developed ectocones. The laterals have well-marked endo- and ectocones, nearly equal in size. Most of the marginals lose the ectocones, but five or six of the innermost teeth retain them. Formula .17.27, with slight variation (see p. 157). Alimentary system The brown oesophagus leads to a long capacious crop, and the intestine (which exhibits marked spiral torsion) extends well behind the stomach, thus contrasting strongly with the condition in Arion. The rectum runs forwards above the common stem of the retractor muscles. Reproductive system (Text-fig. 8 A) The spiral ovotestis, mainly hidden by the upper lobes of the liver, gives off the pale hermaphrodite duct. This is much folded in its anterior wider part, and the terminal slender portion bears an oval seminal vesicle at its entry into the linguiform albumen gland. The common duct is long, voluminous and folded. The spermatheca is an oval sac, with a short duct arising from the oviduct 1-5 mm before its termina- tion. The prostate is a yellow ribbon-like organ on the common duct. At the beginning of the free oviduct it gives off the somewhat convoluted vas deferens, which enters the apex of the epiphallus. The epiphallus is widest apically and has a rounded eminence or blunt caecum below the apex, a feature that is not shared by M. sowerbyi or M . budapestensis. At the insertion of the retractor muscle, the epiphallus, without any external sign, passes into the penis, which narrows towards its termination in the atrium. Before its termination the penis shows two more or less distinct dilata- tions, the upper one containing the blunt perforate penial papilla with a small pigmented diverticulum. The mucous membrane of the epiphallus is thrown into transverse circular folds, and that of the penis into longitudinal folds. A lobulated gland that can be unravelled into long glistening white tubules opens into the atrium. The atrium lodges a fleshy conical stimulator attached by its larger end, with the tip often folded back. The stimulator is generally smooth but occasionally has two or three minute papillae near the tip. 154 BRITISH SLUGS B row FIG. 9. Representative radular teeth (x 330). A. M.gagates. B. M. sowerbyi. C. M. budapestensis. M. sowerbyi, an individual one day old. D. Representative radular teeth (x 1,000). E. Jaw (x 67). BRITISH SLUGS 155 Spermatophore (Text-fig. 8 E) A brown chitinous tube about 10 mm long, widest at the centre and narrowing at the anterior end. Except at two places on the concave side, it bears prominent spines throughout. The spines have four main branches, and fourteen or sixteen terminal points. Mating Observations are very scanty, but apparently the partners during coitus are in close contact, and little or nothing of the genitalia is visible. Though the stimulator is often found protuded in spirit specimens, it is not visible during coitus. Mating has been recorded in spring, summer and winter. Taylor describes the eggs as 2 X 1-5 mm, transparent and thin shelled. This, if correct, is much smaller than the eggs of M. sowerbyi and M. budapestensis. Ecology Though widely distributed in the British Isles, M. gagates seems to be commoner near the sea than inland, and especially common in the south-west. In parts of Devon it is abundant in root crops, and can be a pest. It inhabits cliffs, wild places and gardens. In captivity, like most other slugs, it eats carrot, potato and oatmeal. Distribution (Map 12) M. gagates probably occurs in all the vice-counties of England, Wales and Ireland, the south of Scotland, the Isle of Man and the Clyde Islands. In Europe, its distribu- tion is essentially north-western, though specimens from Marseilles have been examined. Other records from the Mediterranean area probably refer to a closely- related species, here distinguished as M. cf. insularis (p. 156). M. gagates has been introduced into South Australia, Hawaii, California, Canary Isles and Tristan d'Acunha. It is not yet clear whether other records of introduc- tions into America, Bermuda, Ascension Island, St. Helena, South Africa, Australia, Tasmania, New Zealand and Polynesia refer to M. gagates or to M . cf. insularis. Fossil record Pleistocene to Recent. Material examined Living specimens from Epsom, Surrey ; Southwell, Nottingham ; Porthcawl, Glamorgan ; Haverfordwest, Pembrokeshire ; St. Agnes, Stilly Isles ; and Majorca. Spirit specimens from Aberdeen, and from Marseilles, France ; Santa Cruz, Teneriffe ; Calbaden Canyon, Puenta Hills, Los Angeles, California ; Kensington suburb, Adelaide, South Australia (coll. B. C. Cotton) ; and Haleakala Maui, Hawaii. 156 BRITISH SLUGS Milax cf. insularis (Lessona & Pollonera, 1882) Amalia insularis Lessona & Pollonera, 1882 : 57 (see 1884 : 103). Sardinia and Sicily. Amalia ichnusae Lessona & Pollonera, 1882 : 60 (see 1884 : 106). Sardinia. Notes If the name gagates is to be reserved for the north-western species with the smooth stimulator (p. 149), the valid name for the present species remains to be determined. Several names whose application is uncertain (some of which are listed on p. 150) are available within the gagates complex. It will require a thorough study of the Mediterranean forms to clarify the biological and nomenclatural situation. First British record This is the first record to be published. Mr. Armitage collected specimens near Bexhill, Sussex, about 1948, and the differences between these and M. gagates were confirmed by Mr. Watson (in litt.). All other specimens examined, from various parts of Britain, have proved to be M. gagates. Diagnostic features Differs from M. gagates in having large papillae on the atrial stimulator. Also, the slug is usually larger, blacker and less smooth, with a darker sole. External appearance As M. gagates, except for the usually larger size, more prominent tubercles, and more intensely black coloration, with the sole not conspicuously paler. External variation It is not at present possible to assign named varieties to this poorly understood species. Shell, pallial organs, pedal gland, retractor muscles, nervous system As in M. gagates, as far as is known. Jaw As in M. gagates, though varying from pale brown to nearly black in the material examined. Radula The teeth closely resemble those of M. gagates in form and number. Formulae of radulae examined were: M. cf. insularis C. 18.28 (Catania, Sicily), .17.28 X 105 BRITISH SLUGS 157 and .19.27 (Malta), .20.30 x 102 (Marseilles) ; M. gagates .15.31 X 100 (Aberdeen), .16.27 (Porthcawl), .17.27 X 96 (California), .19.27 X 102 (Marseilles). Alimentary system As in M. gagates. Reproductive system (Text-fig. 8 D) Here the differences from M. gagates are obvious. The stimulator (Text-fig. 8 H), instead of being smooth or having at the most two or three minute papillae near the tip, has large papillae disposed in four longitudinal rows near the base, diminishing to two rows and becoming smaller towards the apex. Sometimes the papillae are mucronate, and in one specimen they were found to be partly retracted. The atrial glands extend further backwards than in M. gagates, in a tongue-like prolongation. The epiphallus does not possess the sub-apical bulge seen in M. gagates. The small internally-pigmented accessory penial papilla, instead of being enclosed in the penis sheath, bulges externally ; but it seems likely that this character varies with the physiological state of the animal, in both species. Whereas Mediterranean specimens examined had four rows of papillae towards the bases of their stimulators, Lessona & Pollonera (1884, pi. 2, figs, i and 2 ; 6 and 7) figure "gagates " and insularis with only two rows. They describe ichnusae as having a single row of papillae, though their figure shows the stimulator as smooth (1884 : 106 and pi. 2, fig. 4). Possibly there is more than one species with a papillate stimulator in the Mediterranean region. Mr. Aten of Zaandam, Holland, reports (in litt.} finding M. gagates and M. cf. insularis in the Pyrenees Orientales, and that the latter when young has only a single row of papillae. This suggests that M. ichnusae may be an immature form. Spermatophore, mating, development, behaviour, ecology Not known. Distribution In Britain, once recorded from Bexhill, Sussex. In Europe, found in the Mediter- ranean region. Occurs together with M. gagates near Marseilles. Fossil record Not distinguishable from M . gagates. Specimens examined Spirit specimens from Marseilles and Valence, Malta, Sicily and Majorca. 158 BRITISH SLUGS Milax sowerbyi (Ferussac, 1823) Limax sowerbyi Ferussac, 1823, 2 : 96, pi. 8. London. Limax carinatus Risso, 1826 : 56. Limax marginatus Jefferies, 1862 : 132. Amalia maculata Collinge, 18956 : 336, pi. 23, fig. 6. Amalia collingei Hesse, 1926 : 139. First British record By Ferussac (1823). Diagnostic features Differs from M . gagates in brown colour speckled with darker patches, pale keel and margin of respiratory orifice, grooves between tubercles pigmented, keel not abruptly truncated, animal less compressed laterally when contracted, mucus yellow and more tenacious. Differs from M. budapestensis in broader, uniformly pale sole. Internally characterized by short blunt atrial stimulator, long conical spermatheca, tapered spermatophore hooked at blunt end and bare of spines at narrow end. External appearance (PI. 2, fig. 23) Length when extended about 7 cm. When contracted the slug is rather less compressed laterally than M. gagates, and the keel is not abruptly truncated behind. The body is usually pale or dark brown with darker speckling, with the keel paler and the grooves between the tubercles pigmented. The mantle grooves are as in M. gagates, but the respiratory orifice is pale margined. The tripartite sole is pale, and the sticky mucus yellow. Variation Var. alba Taylor (1904), entirely white ; var. fuscocarinata Cockerell (i886c), keel the same colour as the back ; var. nigrescens Cockerell, black with orange keel ; var. oretea Lessona & Pollonera (1882), with median dark streak on mantle found in Devon and Glamorgan. Shell (Text-fig. 8 i) Larger, thicker and more evenly oval than that of M. gagates, measuring 5x3 mm. Pallial organs As in M. gagates. Pedal gland The duct does not project beyond the glandular tissue. Retractor muscles The pharyngeal retractor is more deeply cleft than in M. gagates, and the two tentacular retractors separate sooner. The penial retractor arises nearer to the BRITISH SLUGS 159 common stem of the other muscles, and there is a pair of short muscles from the penis to the body-wall. Nervous system As in M. gagates. Jaw As in M. gagates, but a little broader from upper to lower margin. Radula (Text-fig. 9 B) As in M . gagates, except that there are a few more transverse rows of teeth, and there are fewer laterals and more marginals in each row, with two or three of the innermost or transitional marginals often retaining a minute ectocone. Formula 0.13.36 x 115. Alimentary system As in M. gagates. Reproductive system (Text-fig. 8 B) The lower genital ducts differ from those of M. gagates in several features. The short wide spermatheca duct arises from the free oviduct even nearer the atrium, and the sac has a narrowing backward prolongation. The epiphallus widens instead of narrowing towards its termination, and does not present the sub-apical lateral bulge of M. gagates, nor the small pigmented diverticulum at the distal end of the penis. The atrium contains a short blunt smooth stimulator. Spermatophore (Text-fig. 8 F) Thicker than in M. gagates, measures about 13 mm long. The thinner anterior end is smooth, and the curved posterior end bears recurved spines of three types. As usually found in the spermatheca, and figured by Taylor (1902-07, fig. 174) and others, the smooth anterior end terminates abruptly. However, if the spermatophore is taken very shortly after coitus, before the spermathecal fluid has had time to act upon it, the anterior end is found to be capped by a curious conical structure about 2 mm long, with a rounded apex, and formed by four leaflets with fringed edges (Quick, 1950). A fresh spermatophore of M. budapestensis does not have this cap, but no opportunity has occurred of examining a fresh spermatophore from M. gagates. Occasionally two spermatophores are found in the spermatheca in M. sowerbyi, but whether they are transferred during a single coitus (which is very prolonged) or on two separate occasions is not known. Mating After following and then curving around, the pair remain in close contact for twelve to eighteen hours, and little or nothing is visible of the genitalia. Pairing has been observed in October and November, but perhaps occurs at other times. 160 BRITISH SLUGS Development The eggs are 4x3-5 mm, soft, translucent and pale amber, laid in clusters of about a dozen, hatching in from four to six weeks. The embryo before hatching has a large flattened caudal vesicle nearly circular in outline. When hatched the young are 8 mm long, with a conspicuous pale keel. The mantle is speckled with black and sometimes has a dark median streak, which is retained in var. oretea when adult. Ecology M. sowerbyi is found in gardens, cultivated fields and rubbish heaps. It burrows in the ground, and can do considerable damage to carrot and potato crops. Distribution (Map 13) Occurs in probably every vice-county of England, Wales, Ireland and the southern half of Scotland, and in France, Belgium and Spain. Probably known under other names from Portugal, Italy, Greece and the north coast of Africa. Introduced into South America and New Zealand? Fossil record Pleistocene to Recent. Material examined Living specimens from many places in Glamorgan, Pembroke, Surrey, Berkshire and Sussex. Spirit material from Britain and Marseilles. Milax budapestensis (Hazay, 1881) Limax gracilis Leydig, 1876 : 276, nee Limax gracilis Rafinesque-Schmaltz, i82oa. Amalia budapestensis Hazay, 1881 : 40. Budapest. First British record By Phillips & Watson (1930). Diagnostic features Sole with dark central and paler lateral areas ; body when contracted tends to lie in a curve and is less compact and humped than M. gagates and M. sowerbyi. No atrial stimulator, spermatophore slender and spirally twisted. Radula has more marginal teeth, and spermathecal duct is longer, than in M. gagates and M. sowerbyi. External appearance (PI. 2, fig. 20) The smallest British species of Milax, very slender when extended to 5 cm. The dorsum is dark, sometimes nearly black or with a brownish tinge, with a dirty BRITISH SLUGS 161 yellow or orange keel. The grooves between the dorsal and the lateral, somewhat polygonal, tubercles are deeply pigmented. The mantle is sometimes paler than the dorsum, the horseshoe-shaped groove deeply pigmented, and the small respiratory orifice has a grey margin. The peripodial groove and margin of the sole are deeply pigmented, the lateral areas of the sole grey, and the median area dark grey or nearly black. The mucus is viscid and colourless, but in the mantle area becomes yellowish after repeated irritation. External variation Variation is confined to the intensity of the general grey colour ; to the tint of the keel, which varies from yellow to brownish-orange ; and to the sole, whose lateral areas are occasionally darker than the central area. SMI (Text-fig. 8 L) Concave or flat below, measuring 3 x 1-25 mm, brownish in colour. Pallial organs As in M. gagates. Pedal gland The duct extends behind the glandular tissue for about 5 mm. Retractor muscles The tentacular and pharyngeal retractors are usually rather more deeply cleft than in M. gagates, less so than in M. sowerbyi. Nervous system As in M. gagates. Jaw Brown and oxygnathous as in the other species but smaller, measuring about 1-5 mm across and 0-5 mm in height. Radula Differs from those of the other species in having the basal plate of the central tooth relatively broader, the aculeate mesocones of the marginals more slender, fewer lateral teeth and more marginals. Formula .7-8.38-40. Alimentary system As in M. gagates. Reproductive system (Text-fig. 8 c) The pale spirally twisted ovotestis has five or six closely approximated lobes, 162 BRITISH SLUGS each composed of numerous follicles. The hermaphrodite duct is slender for the greater part of its length, but becomes dilated and folded near the albumen gland, and here shows a small seminal vesicle. The common duct with a broad yellow prostatic ribbon is folded S-wise, and after parting from the vas deferens contracts to form a relatively long slender free oviduct. At about two-thirds of its length, the wider relatively long spermatheca duct arises, leading to a cylindrical sac with a rounded apex. The rather long flexuous vas deferens enters the epiphallus sub- apically. In the writer's experience the epiphallus and penis are long and slender, but Phillips & Watson (1930) describe and figure a short blunt form in addition, and find this the commoner. It is not known if this is a true dimorphism, or only the result of the physiological condition. Hazay (1881, pi. i) shows an intermediate form of penis and epiphallus, and Simroth (1885, pi. 10, fig. 13) the long form. Both forms occur in Belgian specimens, sent to me by Dr. W. Adam. The atrium receives the slender ducts of a pair of lobulated atrial glands, and has no stimulator. Spermatophore (Text-fig. 8 G) Length about 16 mm, anterior end slender, posterior end spirally twisted. No trace of a frilled cap as in M. sowerbyi, even in sperm at ophores removed during or immediately after coitus. Except for the first 2 mm at the anterior end, the Spermato- phore is covered by forwardly-recurved branching spines, which are smaller and less branched towards the ends. There are two rows of spines anteriorly, three or four rows on the posterior two-thirds. Mating Occurs during November, December and January, under cover of loose stones or logs. Coitus generally starts in the evening and continues until the following mid-day or later. A little of the partly everted atria and the two penes is visible between the necks of the partners. The atrial glands are partly contained in the everted atria, together with the distal ends of the genital ducts. Development The ellipsoidal yellow leathery eggs are laid in small clusters underground, and measure 2-9 X 2-25 mm. Those laid in December hatch in March. The young at hatching are 4-5 mm long, pale yellowish-grey with a recognizable keel. Behaviour Whereas when at rest the other species of Milax assume a compressed helmet shape, M. budapestensis is usually found only half contracted, and curved into a semicircle. Ecology Occurs mostly in gardens and allotments, under stones and rubbish, and can do considerable damage to roots and tubers. Though much less common in wild places, there are records for such situations in Wales and Cheshire. BRITISH SLUGS 163 Distribution (Map 14) Recorded from many localities in the British Isles, mainly in western England and Wales, but also in south-eastern England, southern Ireland and the Hebrides. On the Continent, known from Belgium, Germany (Wurtemberg, Tubingen, the Rho'n Mountains and Wiirzburg), North Italy (Verona, Vicenza and Padua), Hungary (Budapest), Transylvania, Bulgaria and Crimea. It probably occurs in many places between these widely-separated localities, in the zone from the British Isles to the Crimea. For a map and further details, see Phillips & Watson (1930) and Ellis (1951). Fossil record Recent. Material examined Living specimens from Berkshire, Glamorgan, Pembroke and Denbighshire. Spirit material from Britain and Belgium. Subfamily LIMACINAE Limacidae in which the keel rarely extends forwards as far as the mantle. The nucleus of the enclosed shell is terminal and lies to the left of the mid-line. The mantle is concentrically ridged and lacks a horseshoe-shaped furrow. There is no well- developed epiphallus, no spermatophore and no atrial gland. The stimulatory organ, when present, is a penial sarcobelum, and not an atrial stimulator as in Milax. The endocones of the lateral radular teeth, when present, are more or less united with the mesocones. The intestine has one or two forwardly-directed loops. Genus AGRIOLIMAX Morch, 1865 Type species Limax agrestis Linnaeus, 1758. The right ocular retractor lies to the left of both penis and vagina. The intestine has one forwardly-directed loop, and the right lobe of the liver forms the apex of the visceral mass. A penial sarcobelum is present, more or less developed. The nucleus of the concentric mantle ridges lies to the right of the middle line, over the respiratory orifice. The tail is obliquely truncate. Pilsbry (1944, 1948) adopts the name Deroceras Rafinesque-Schmaltz (18200 : 10) for this genus, as do most American and many European authors. Watson (1943 : 54) discusses the question. Ellis (in press) has submitted to the International Commission on Zoological Nomenclature a proposal for the suppression of Deroceras and other older names, and the validation of Agriolimax. Hesse (1926) mentions sectional names for the first three species considered here. Agriolimax s.s. (including A. agrestis and A. reticulatus) has a rectal caecum and 164 BRITISH SLUGS a more or less tufted penial appendix, whereas Hydrolimax Malm (including A . laevis) lacks both caecum and appendix, but has a long tubular penial diverticulum. Other sectional names have been proposed for groups of extra-limital species, such as Krynickillus Kaleniczenco, which lacks caecum, appendix and sarcobelum. However, several species show combinations of characters not covered by existing sectional names. A. caruanae has a tufted penial appendix like Agriolimax s.s., but lacks a rectal caecum like Hydrolimax ; while one or both of its penial diverticula may correspond to the latter's single but minutely bifid diverticulum. A slug collected in Cyprus by A. R. Waterston (? A. cyprius Simroth, 1906, or A. panormitanus Lessona & Pollonera, 1882) has a tufted appendix and a single diverticulum. Of two Abyssinian slugs described by Quick (1954) , A . gughensis lacks caecum, appendix and diverticula but has a long, narrow, backwardly-directed sarcobelum ; while A, scotti lacks caecum, appendix and sarcobelum, but has two diverticula. In view of the intergradation and overlapping between species of which the anatomy is known, and of the number of species (especially in south-eastern Europe) of which it is not known, the use of sectional names within this genus does not seem useful. Agriolimax reticulatus (Miiller, 1774) Limax reticulatus Miiller, 1774 : 10. Denmark. Limax canariensis d'Orbigny, 1839 : 47. Krynickillus minutus Kalenickzenko, 1851 : 224. Notes The name agrestis has been used for this species by most British authors (see p. 170). First British record By Lister (1674 : 99) . Diagnostic features Distinguished from A. laevis and A. caruanae by larger size, opaque appearance with very variable coloration, white sticky mucus and less active movements ; and internally by the presence of a rectal caecum. Differs from A . agrestis in the trifid crenulate penial appendix, and the posterior situation of the ovotestis. External appearance (PI. 2, fig. 15) Length about 3-5 cm. The tail is obliquely truncated, the back keeled for a short distance at the hind end, and the sole tripartite. The respiratory orifice is at about two-thirds of the distance from front to back of the mantle, and the reproductive orifice close behind and below the right upper tentacle. The concentric striae of the mantle disappear in alcohol. The mucus is whitish on account of calcareous granules. BRITISH SLUGS 165 The colour is variable, with the body usually marked with brown or grey. The respiratory orifice has a conspicuous pale border. External variation The colour varies from pale cream to bluish-black, but is usually some shade of brownish-yellow, with darker grooves and scattered dark brown markings of varied shapes and sizes. Named varieties and subvarieties are very numerous. Many are practically synonymous, while some of those listed by Taylor (1902-07) refer to other species. Individuals often become darker as winter approaches. The following varieties may be recognized : var. cineracea Moquin-Tandon (1855), entirely ash coloured ; var. violacea Gassies (1849), lilac or slate-coloured ; var. rufescens Dumont & Mortillet (1856), rufous ; var. reticulata Miiller (1774), grooves dark, giving a reticulate appearance ; var. nigra Morelet (1845), black, with pigment sometimes invading the sole. Albino forms occur occasionally. Shell (Text-fig. 10 F) The right border is convex. Pallial organs (Text-fig. 10 A) When the mantle is reflected to the right, a lobe of the kidney is seen to cover the rectum, and the short retractor penis to arise from the diaphragm at the anterior border of the pericardium. Retractor muscles (Text-figs, n G and H) The cephalic retractor muscles arise as a single stem in the mid-line, a short distance behind the diaphragm. The main stem usually divides basally, to form the right ocular retractor, and a common stem for the left ocular and the buccal bulb retractors (Text-fig, u G), but variations occur (Text-fig, n H). The genital retractor arises from the diaphragm near the front of the kidney, and is inserted between the vas deferens and the constriction above the sarcobelum. Nervous system This is very similar to that of Milax. Jaw Pale golden, of the usual oxygnathous type. Radula (Text-figs, n L and o) As in all Agriolimax, the endocones of the lateral teeth are delicate, lying close to the mesocone and on a slightly more superficial plane. The basal plates of the central and lateral teeth are larger and relatively narrower than those of A. agrestis. Table I compares the radulae of the four British species of Agriolimax. 166 FIG. 10. Agriolimax. Dissection with mantle reflected to the right. A. A. reticu- latus (x 2), ovotestis extends much further backwards than in A. agrestis. B. A. agrestis (x 2), ovotestis does not extend much further back than the rectal caecum. C. A. caruanae (x 2-7), ovotestis far forwards, exposed on left side. No rectal caecum. D. A. laevis (x 2-7). A. caruanae. E. Pallial region and cephalic retractors (x 12). K. Mating pair with sarcobela protruded. Shell (x 4). F. A. reticulatus. G. A. agrestis. H. A. caruanae. I. A. laevis. BRITISH SLUGS 167 TABLE I. Radular Formulae of the British Species of Agriolimax Morch Locality Date Formula (mm) Remarks Capel-y-mn, Brecknock Blackpill, Swansea Mumbles Rd., Swansea Porteynon, Glamorgan Clyne Valley, Swansea Llanwrtyd Wells Burpham, Surrey Wheatfen Broad Singleton, Swansea . Sketty, Swansea Singleton, Swansea . Porteynon, Glamorgan Date Formula Central tooth (mm) Agriolimax reticulatus (Miiller) 2o.vii.46 20 . vii . 46 17. viii. 46 . .14.22x86 .17.27x107 . .14.22x88 0-043x0-022 ? 23 . viii . 46 . 7 . ix . 46 i8.xi.38 . .14.21x100 .16.22 .17.26x100 .6.8 0-035 xo-oi7 0-049 x 0-024 0-043x0-022 29.xii.46 . .18.26x122 . Agriolimax agrestis (Linnaeus) 0-052x0-026 Agriolimax laevis (Miiller) io.iii.48 . .13.27x92 . .13.25x95 2 . viii . . Agriolimax caruanae Pollonera ? .x.3i . .14.50 ? .iv.32 . .14.55 ? .vii.46 C.io.? .14.55x118 0-023x0-012 0-023x0-012 0-023 xo-oi4 0-029x0-017 Small but mature. Young, 8 mm long. 17. viii. 46 . .16.23 0-038x0-019 Radula 3-2 x 1-7 mm 23 . viii . 46 . C.i6.23x 105 . 040 X 02 . 17. viii. 46 . .14.21 x 100 . 0-034x0-017 . ? .x. 4 6 . .15.24 0-038x0-019 . ? . ix . 46 .15.? 0-035x0-017 . ? .15.21 x ioo . 0-032x0-016 . Euphallic. Hemiphallic. Young, 12 mm. Radula 3-1 x 1-5 mm. Alimentary system (Text-fig. 10 A) The tabulated salivary glands lie on each side of the crop, which narrows as it enters the small stomach. The posterior loop of the intestine extends behind the stomach towards the apex of the visceral mass, which is formed by the right lobe of the liver. The rectum has a small caecum overlying the ovotestis. Reproductive system (Text-fig, n c) The brown tabulated ovotestis lies towards the rear of the visceral mass, and is partly exposed on the surface (Text-fig. 10 A). The hermaphrodite duct is short and only slightly sinuous, with a seminal vesicle at its slender termination where it enters the albumen gland. The latter is short, shaped like a mammalian liver, and ZOOL. 6, 3. 12 1 68 BRITISH SLUGS FIG. ii. Agriolimax. Distal genital ducts (x 5-3). A. A. latvis euphallic indi- vidual. B. A. laevis aphallic individual. C. A. reticulatus. D. A. agrestis. E. A. laevis apex of penial appendage, showing bifid tip and two pilasters. Cephalic retractor muscles. F. A . laevis. G & H. A . reticulatus. I & K. A. agrestis. Representative radular teeth. L. A. reticulatus (x 330). M. A. agrestis (x 330). N. A. laevis (x 330). O. A. reticulatus (x 500) from young animal 8 mm long. BRITISH SLUGS 169 pale brown in colour. The common duct, with the lobulated prostatic ribbon, is thrown into zig-zag folds. The free oviduct is short and straight, and the atrium small. The spermatheca duct arises from the atrium and expands to a narrow oval sac. The short vas deferens enters the penis beyond its apex, beneath the genital retractor muscle. The penis is narrow basally, expands to a segment containing the triangular sarcobelum, and narrows a little before the apical segment. The latter terminates in a large penial appendage which is trifid, each branch having a crenulated outline. The details of the branching vary to some extent. The branches, which are tubular, are everted and turned completely inside out during mating, as in A. agrestis and A. caruanae (see Text-figs. 12 D and E). Mating This can occur throughout the year in Britain, even in the winter months, taking place on the surface usually after dark. The pair follow one another for from half to one hour in a narrowing circle, with copious exudation of mucus. The sarcobelum is protruded and plays over the partner's body. Then the animals remain still, neck to neck with bodies curved, and each atrium and penis is everted, bringing the oviduct and spermatheca duct to the surface. Then the base of the penial appendix appears and is explosively turned inside out, the branches appearing as long thin hyaline threads. A mass of sperms enclosed in a ball of mucus is simul- taneously transferred from one partner to the other's everted mass, to which it adheres. The animals separate at once, and in a few seconds the genital complexes are invaginated and withdrawn, carrying the sperm packets with them. Development The eggs, laid about ten days after mating, are about 3 X 2-5 mm, translucent and speckled with calcareous dots which become absorbed during development. They hatch in two or three weeks, the young being 4 mm long, pale grey and trans- lucent, and like all the British species have no bands at any stage. The young grow rapidly, and in warm weather may mature in three months. During late embryonic development the caudal and cephalic vesicles can be seen pulsating through the translucent shell, when the eggs are viewed under water (see Text-fig. 18 G). In Britain several broods are produced by a pair, and 700 eggs may be produced in a season. Luther (1915) states that self-fertilization occurs in A . reticulatus isolated from birth. Ecology A. reticulatus is perhaps the commonest British slug. It prefers drier sites than A . agrestis and occurs in grass fields, root crops, hedges, gardens and rubbish heaps, and under stones and logs in woods and coppices. It is often a serious pest in gardens and allotments, devouring seedlings and green crops. Distribution (Map 15) Ubiquitous in the British Isles. Probably indigenous to most of Europe, the 170 BRITISH SLUGS Mediterranean region and the Atlantic isles, and introduced by commerce into most parts of the world. Fossil record Pleistocene to Recent. Material examined Living specimens from many places in England and Wales. Spirit material from Britain, Europe, Atlantic isles, Reunion, America and New Zealand. Agriolimax agrestis (Linnaeus, 1758) Limax agrestis Linnaeus, 1758 : 652. Sweden. Limax bilobatus Ferussac, 1819 : 74. Limax pallidus Schrenk, 1848 : 143. Notes The commonest British slug, A. reticulatus (Miiller), was formerly regarded by British authors as a variety of A. agrestis, though Luther (1915) had shown them to be distinct. It is uncertain which of many synonyms in this group of slugs belong to A. agrestis and which to A. reticulatus. First British record By Ellis (1941). The slug was found at Wheatfen Broad, Norfolk, and determined by H. Watson. Diagnostic features Agrees externally with A. reticulatus, except that it is smaller, smoother and more slender and seldom has more than a few small dark flecks. Internally, the ovotestis lies further forward, and the penial appendix is short, smooth and unbranched. External appearance Like A. reticulatus, but rather slimmer and smoother, and less pigmented. The body is pale greyish-yellow, paler on the sides, sometimes with whitish calcareous- looking spots. External variation Occasionally, a few dark flecks are present and the grooves are slightly pigmented. Shell (Text-fig. 10 G) About 3 mm long, usually with the right border slightly concave. BRITISH SLUGS 171 P allied organs (Text-fig. 10 B), retractor muscles (Text-figs, n I and K), nervous system, jaw As in A. reticulatus. Radula (Text-fig, n M) The teeth are a little smaller than in A. reticulatus, and the basal plate wider (Table I, p. 167). Alimentary system (Text-fig. 10 B) As in A. reticulatus. Reproductive system (Text-fig, n D) The ovotestis lies further forward than in A. reticulatus, near the centre of the visceral mass, and is only partly exposed on the right side beneath the rectal caecum (Text-fig. 10 B). The remainder of the reproductive system is like that of A . reticulatus, except for the apex of the penis, which terminates in a single small curved digitiform appendix. Mating As in A. reticulatus. Mating occurs in Britain in the autumn, and perhaps at other times. Luther (1915) states that this species is incapable of self-fertilization, unlike A. reticulatus. Development The eggs are laid in clusters of from ten to twenty or more. They measure from 2-2 X 2-0 to 2-5 X 2-25 mm, and are more translucent and more finely dotted with calcareous particles than those of A. reticulatus. The newly hatched young are 3-5 mm long, translucent and whitish, and have no bands at any stage. In Britain, the eggs are laid in the autumn and hatch in three or four weeks. The young mature in the following summer and autumn, and die in the late autumn after mating. In Finland, the life cycle is different (Luther, 1915), in relation to the more severe climate. The eggs are laid, and the parents die, in late August or September, while the eggs do not hatch until the following June. Thus embryonic life lasts for more than eight months in Finland and less than a month in England, while post-embryonic life lasts about three and twelve months respectively. Ecology In Britain, apparently confined to marshy carr (see Ellis, 1941) in East Anglia. Distribution (Map 16) In Britain, certainly recorded only from marshes by the Norfolk Broads : a single record of an immature specimen from North Uist (Ellis, 1951 : 196) has not been confirmed. Probably widespread in Europe, northwards to Finmark and Iceland, 172 BRITISH SLUGS though the detailed distribution is uncertain because of confusion between this species and A. reticulatus. For example, Germain (1930) describes and figures A. reticulatus under the name agrestis. Material examined Living specimens from Wheatfen Broad, Sutton Broad and Alderfen Broad, all in east Norfolk. Agriolimax laevis (Miiller, 1774) Limax laevis Miiller, 1774 : i. Frederiksdal, Denmark. Limax brunneus Draparnaud, 1801 : 104. Limax campestris Binney, 1844 : 52. Limax hyperboreus Wester lund, 1876 : 97. Agriolimax bevenoti Collinge, 18970 : 295. Notes Miiller described his slug as " totum nigrum ", except for the median area of the sole, and compared its general appearance with that of the black land planarian Rhynchodemus terrestris. However, topotypes from Frederiksdal agree with the common slug, known as A. laevis in Britain, Europe and North America, in being some shade of brown. The discrepancy in colour is not highly significant, since the present species becomes very dark when somewhat dry and contracted, and since in America it varies in colour from amber to black (Pilsbry, 1948 : 540, 548 ; Altena, 1958 : 30). Simroth (1885 : 222, pi. 9) described and figured, as forms of A. laevis from Germany, what are evidently representatives of two distinct species : those corre- sponding to the euphallic and aphallic forms of the widespread slug (his figs. 21 G and 22 H), and a larger and darker species with a hammer-headed penis (his figs. 17 c and 18 D). The latter is found principally in Germany and Poland, though it appears to have been introduced to Gotland Island, Sweden (Lohmander, in Hit.}. It has never been recorded from Denmark and Lohmander, despite his considerable experience of Danish slugs, has never found it there. Taylor's figures (1902-07, figs. 131 and 132), purporting to show the genitalia of British A. laevis, are copied from Simroth and show this second species. It seems that the name laevis can properly be retained for the species to which it has long been applied. There is no reason to transfer it to the species with the hammer-headed penis, the correct name for which is uncertain. First British record By Johnston (1838 : 154), as Limax brunneus. Diagnostic features Small size, colour light or dark brown, smooth, active. Much smaller than A. caruanae. Head and neck extend further from the mantle than in A. reticulatus. Internally distinguished from the other species by its more sinuous penis, with a BRITISH SLUGS 173 tubular caecum or appendix, and the frequent occurrence of aphallic specimens ; and from A . reticulatus and A . agrestis by the absence of a rectal caecum. External appearance (PI. 2, fig. 19) Length 1-7-2-25 cm. Light or dark brown, translucent, sometimes greenish, more or less flecked with darker spots. The respiratory orifice is less conspicuously pale- bordered than in the other species. The head and neck are thrust further forward than in A. reticulatus and A. agrestis, so that the mantle is more centrally placed, and the mantle has fewer concentric ridges. The mucus is clear, not milky. The slug is very active, and crawls rapidly. External variation The ground colour varies from pale to dark brown, with more or less darker flecking on body and mantle. Pale and dark specimens occur together at all seasons. Named varieties are var. lacustris Bonelli (in Lessona & Pollonera, 1882) ( var. maculata Cockerell, 18866), irregularly spotted with dark brown ; var. grisea Taylor (1904), grey or greenish-grey. There is also marked genital polymorphism, with euphallic, hemiphallic and aphallic individuals. Shell (Text-fig. 10 i) Relatively short and broad. Pallial organs (Text-fig. 10 D) As in A . reticulatus. Retractor muscles (Text-fig, n F) The cephallic retractor muscles arise by a long common stem, and so are less deeply divided than in the other species. The retractor penis is occasionally absent. Nervous system As in A . reticulatus. Jaw Pale golden, i mm across. Radula (Text-fig. UN) Rather fewer lateral teeth, and small ectocones on more of the external marginals, than in A. reticulatus and A. agrestis. Formula .13.26. Alimentary system The rectum lacks a caecal diverticulum, as in A . caruanae. The visceral cavity is often somewhat pigmented, though less so than in A . caruanae. 174 BRITISH SLUGS Reproductive system (Text-figs, n A and B) The dark ovotestis is less exposed than in the other species. The hermaphrodite duct is short and nearly straight, with the usual seminal vesicle at the short brown albumen gland. The female portion of the common duct is leaden grey, and the prostate yellow. The spermatheca is normally oval as in the other species, but sometimes in young examples it is globular. The penis in a euphallic individual shows four regions : a distal, narrower segment ; a wider one enclosing the triangular sarcobelum ; a twisted and contorted segment entered distally by the vas deferens, and giving attachment to the retractor penis ; and a short narrower segment (diverticulum or appendix), more or less bifid at the tip. The appendix contains two thickened pilasters, which fade out on reaching the sarcobelar segment. In the aphallic form (Text-fig, n B) prostate and spermatozoa are present, but the vas deferens fades out or ends in a little vesicle, before reaching a minute knob represent- ing the penis. Apparently the penis is always the last part to be fully developed ; but in young examples that will become euphallic the vas deferens does reach the small penis and a retractor muscle is often present. Stages can be found between this and the adult euphallic form. Babor (1894) records observations which he thought indicated that A. laevis was normally protogynous young adults being always female and middle-aged adults hermaphrodite, with some old animals attaining a purely male phase, with atrophied albumen gland and spermatheca and hyper- trophied penis. He states that in the young stages sperms are very few, and there is no sperm duct or penis. However, dissection of many British examples, taken at various sites and different seasons, shows that here sperms, prostate and vas deferens are always present ; though in animals that will eventually be aphallic the vas deferens ends blindly. No example has been seen in which the albumen gland and spermatheca have atrophied. As these forms often occur together, in varying pro- portions in the same site at different seasons, it seems more likely that this is a case of genital dimorphism, such as occurs in Zonitoides. To establish this with certainty it is desirable that samples should be taken from the same site each month for a year, but an opportunity to do this has not occurred. Pilsbry (1948 : 539-552) gives numerous figures of similar dimorphism in the American forms of A . laevis. Mating Gerhardt (1939) says that following and circling around is omitted in A. laevis, and that the partners remain in contact before the sperms are transferred for a much longer period (up to an hour) than in the other species. Development The eggs are laid singly or in small clusters, probably at all seasons of the year, having been found in January, April and October. They are relatively large, from 1-8 x 1-5 to 2 X 1-3 mm, translucent, resilient, finely punctate, with polygonal calcareous granules. They hatch in four or five weeks. The young are 4 mm long, white and translucent, with the opaque white liver lobes showing through the integument. The head and tentacles are light pinkish-brown with a violet tinge. BRITISH SLUGS 175 After a few days, pigmentation begins at the tail and head, and gradually invades the rest of the body. Ecology A. laevis inhabits damper sites than our other slugs and is common in marshes and on the borders of ponds and rivers, but is occasionally found in drier sites, in fields and on the borders of woods. It has been seen feeding on mealy bugs infesting Streptocarpus in greenhouses. Distribution (Map 17) Probably ubiquitous in the British Isles. Its continental distribution is somewhat uncertain, because of the existence of a somewhat similar slug distinguished by a hammer-headed penis (p. 172). It is common in marshy places throughout the Holarctic, extending southwards in Europe to northern Italy. Most of this range is probably occupied by the true A . laevis, showing only slight infra-specific variation. A. laevis has been widely introduced into tropical regions and oceanic islands, where it has been recorded under many different names. For example, specimens from Hawaii in the B.M. (N.H.) labelled as A. bevenoti Collinge (Kauai) appear to belong to the aphallic form of this species, and others labelled as A . perkinsi Collinge (Lanai) and A . globosus Collinge (Mauna Loa) to the euphallic form. Fossil record Pleistocene to Recent. Material examined Living specimens from many places in England and Wales. Spirit specimens from the type locality (Frederiksdal, Denmark), kindly collected and sent to me by Dr. Mandahl-Barth, and from Britain, South America, New Zealand and Hawaii. Agriolimax caruanae Pollonera, 1891 Agriolimax caruanae Pollonera, 1891 : 3. Porta Reale, Malta. Notes Pollonera described the keel of his species as blackish, and figured a more globular spermatheca than is found in adult British specimens. However, in Britain the keel and head are the first parts to pigment, so that in young animals they are darker than the rest of the body. The spermatheca is sometimes globular, in young animals at any rate in Britain, and in adults of some American forms. Pilsbry (1948 : 558) figures considerable variation between American specimens, which he nevertheless recognizes as belonging to A . caruanae. The British slug can be at most a geographical race of this species, of which Quick (1949 : 28) gives a short account. I 7 6 BRITISH SLUGS FIG. 12. BRITISH SLUGS 177 First British record Specimens, found by the late Charles Oldham in Cornwall about 1930, were identified with this species (as perhaps a distinct subspecies or variety) by H. Watson, and exhibited at meetings of the Conchological Society of Great Britain and Ireland and the Malacological Society of London. Diagnostic features Distinguished in life from the externally similar A . laevis by larger size and even greater activity. Internally distinguished by having two penial diverticula and a tuft of appendages, a darkly-pigmented visceral cavity, and a greater number of marginal radular teeth. External appearance (PL 2, fig. 17) The length when extended is 2-5-3 cm > r a little more. As in A. laevis, the head and neck extend far beyond the mantle. The slug is dark translucent chestnut brown, greyish-brown or grey, with the mantle lighter over the pallial region, and the body and mantle more or less flecked with dark brown. The respiratory orifice is pale rimmed. The sole is grey rather than brown, because the dark visceral pig- mentation shows through the tissues. The mucus is thin and colourless. External variation The colour varies from a clear chestnut-brown to a greyer brown, and the small darker flecks on body and mantle are more or less marked, but there are no named varieties. Shell (Text-fig. 10 H) 4-5 X 2 mm, or a little larger. In undistorted specimens, the right border is slightly concave. Pallial organs (Text-fig. 10 E) As in A . reticulatus. FIG. 12. Agriolimax caruanae. A. Reproductive organs (x 4). B. Penis ( X 14) opened to show sarcobelum, two diverticula, entries of vas deferens and crenu- late branched appendix, penial retractor. C. Penis (x 8). D. Sarcobelum protruded, atrium everted, appendix beginning to be everted. E. Three branches of appendix fully everted, sarcobelum beginning to shrink. F. Distal genital ducts ( x 30) of young individual 1 1 cm long, showing oviducts, spermatheca, sarco- belum in penis, rudiments of diverticula and appendix. G & H. Cephalic retrac- tors. The buccal retractor is sometimes completely divided. I. Ganglia and commissures of nerve ring. K. Jaw (x 20). L. Representative radular teeth (x 500). M. Jaw ( x 100) of individual one day old. N. Jaw ( x 100) of young individual 6 mm long. 178 BRITISH SLUGS Retractor muscles (Text-figs. 12 G and H) The cephalic retractors arise by a common stem, and the buccal retractor is divided either completely or for half its length. Nervous system (Text-fig. 12 i) As in A . reticulatus. Jaw (Text-fig. 12 K) Of the usual oxygnathous type, 1-3 mm across. In the newly-hatched animal, a median indentation (like that seen in Milax, Limax and Vitrina] reveals a bilateral origin (Text-figs. 12 M and N). Radula (Text-fig. 12 L) Characterized by having rather fewer lateral teeth and more marginals than A. reticulatus and A. agrestis. Formula 0.14.55. Alimentary system (Text-fig. 10 c) As in A. laevis, there is no rectal caecum. The walls of the visceral cavity are darkly pigmented. Reproductive system (Text-figs. 10 c and 12 A-F) The dark ovotestis lies rather far forwards, and is partly exposed on the left side of the visceral mass. The oviduct, atrium and spermatheca require no special comment. The basal segment of the penis is slender and the segment containing the sarcobelum capacious. The distal segment bears apically two curved hollow processes, of which the left is often the longer, and between these an appendix with four or five long slender crenulated branches, corresponding with the apical appendix of A. reticulatus. Mating (Text-fig. 10 K) Breeding occurs throughout the year, eggs having been seen in January, May, June, July, October, November and December. A mating pair was observed on 25th June, 1950. At 6.10 p.m. the partners were following one another in a circle, with protruded sarcobela, licking one another's tails. At 6.40 they were lying quietly with their necks close together, with the atria everted and the sarcobela shrinking. A few seconds later, the four long tubular branches of the penial appendices were everted, lying across the partner's body like hyaline threads. At 6.42 the everted organs were rapidly withdrawn, with the sarcobela last, and the animals crawled apart. Individuals from Glamorgan and Denbigh, isolated from birth, have on several occasions been seen to lay a few self-fertilized eggs, which have developed to an advanced stage although they have not been seen to hatch. Gregg (1944) reports fertile eggs from virgin specimens in California. There this species (for which Gregg BRITISH SLUGS 179 uses the name Deroceras panormitanus] apparently passes through a purely female phase when about half grown. Only about half the individuals survive this phase, to grow to full size and develop male organs. There is no sign of such a female phase in Britain, where slugs only 10 mm long, with their common ducts still slender and immature, always have penes in which the rudiments of all the adult features are easily recognizable. Two diverticula, a budding tufted appendix and a sarcobelum can be seen in a penis between 1-3 and 1-6 mm long. Aphallic and hemiphallic individuals are unknown in Britain. In any case, aphallic individuals of A. laevis are not in a female phase, since they have spermatozoa and a prostate. Gregg's observations are explicable in two ways : he may possibly have been working with a mixed culture of A . caruanae and A . laevis, or the difference may be related to the more rapid development of A. caruanae in California than in Britain. Development The eggs are relatively small, from 1-5 x 1-5 to 175 x 1-4 mm : much smaller than those of A . reticulatus and A . agrestis, and not much larger than those of A . laevis. They are laid in clusters of up to fifty eggs, and in the summer hatch in about seven- teen days. The young are only about 3 mm long, very pale and translucent, with pale violet tentacles. Pigmentation begins at the head and tail, as in A . laevis. Behaviour A . caruanae crawls extraordinarily fast for a slug, and is irritable and pugnacious, snapping at its neighbours and lashing its tail. Even when well supplied with food, it shows cannibalistic tendencies. Ecology Usually confined to gardens, in flower and vegetable beds, lawns, rubbish heaps and greenhouses ; but occasionally found at a distance from houses, in hedges, fields and waste places (records from Devon and Glamorgan). Distribution (Map 18) Recorded from scattered localities in England (mainly in the west), Wales, Scotland and southern Ireland (Makings, 1959). The extra-limital distribution is very imper- fectly known. Repeated efforts to obtain topotypes from Malta have failed. The species probably occurs at Marseilles (Pilsbry, 1948 : 560), and has been introduced into California. Fossil record Holocene shells from Romney Marsh (Hayward, 1954^ and b) are very suggestive of this species. Although slug shells are badly characterized, especially when fossil, there is a prima facie case for believing that these are indeed shells of A. caruanae. Otherwise, the species is known only from Recent deposits. Material examined Living or spirit specimens from St. Agnes, Scilly Isles ; Exeter, Devon ; Swansea i8o BRITISH SLUGS and Porteynon, Glamorgan ; Old Colwyn, Denbighshire ; St. Albans, Hertfordshire ; Bromborough, Cheshire ; Edinburgh and Aberdeen, Scotland ; Cork (grounds of the University, from Dr. Makings) and Newcastle, Co. Down (Dr. Stelfox's garden), Ireland ; Pyrenees Orientales ; Golden Gate Park, San Francisco. Genus LIMAX Linnaeus, 1758 Type species Limax maximus Linnaeus, 1758. Limacinae in which the right ocular retractor passes forwards between the penis and the vagina. The intestine has two forwardly-directed loops, and the left lobe of the liver forms the apex of the visceral mass. There is no penial stimulator. The nucleus of the concentric mantle ridges lies in the mid-dorsal line. The dorsal keel slopes evenly to the end of the tail, which is therefore pointed and not truncated. Hesse (1926 : 8-16, 76-92) divides Limax into several subgenera, and the nominate subgenus into sections. Since no regular trends in several characters are discernible, each species being specialized in some respect or another (e.g. the long penis and peculiar mating behaviour of L. maximus and L. cinereoniger) , the infrageneric arrangement depends largely on the relative systematic importance attached to the radula, penis, alimentary tract and other organs. For example, L. nyctelius may be associated with L. flavus in the section Limacus, on the basis of its rather long cylindrical penis and long rectal caecum ; but their radulae are very different, that of L. flavus being specialized in the suppression of ectocones and endocones, while that of L. nyctelius approaches the condition of Agriolimax. Subgeneric and sectional names are therefore not adopted here. Limax tenellus Muller, 1774 Limax tenellus Muller, 1774 : n. Denmark. Limax tenellus; Nilsson, 1822 : n. Limax serotinus Schrank, 1848 : 144. Limax cereus Held, 1849 : 15. Limax fulvus Normand, 1852 : 7. Agriolimax tenellus', Lessona & Pollonera, 1882 : 45, pi. i. fig. 7. Notes Although Nilsson is often quoted as the author of this name, it was first applied by Muller. Ignoring an obvious mistake in measurement, Miiller's description seems adequate to identify the present species : " Limax virescens, capite tentaculisque nigris. Dan. SP^ED-SNEGLEN. long. 10 unc. Totus albidus, Clypeus in luteum, abdomen in virescentem colorem aliquantum vergit ; ille margine postico, hoc apice supra nigricat. In Fossulis Nemorum foliis aridis repletis ; primo vere." Altena (1958) confirms that L. fulvus Normand is indeed L. tenellus. First British record By Alder (1848 : 124). BRITISH SLUGS 181 Diagnostic features Distinguished externally from other British species by small size, and pale yellow colour, with black head and tentacles ; and internally by a short penis with no appendix. External appearance (PI. 2, fig. 16) The smallest British Limax, from 25 to 35 mm long when extended. Colour pale yellow, or sometimes reddish-brown, with blackish head and tentacles, and sometimes with a pair of more or less distinct body and mantle bands. The mucus is yellow. External variation Var. cerea Held (1849), uniform yellow, or with mere traces of bands ; var. fulva Normand (1852), reddish-brown, suffused dorsally with black ; var. cincta Heyne- mann (1861), with distinct dark mantle bands, and sometimes body bands. Shell Thin and delicate, measuring about 3-5 x 2 mm. Pallial organs (Text-fig. 13 D) The somewhat angularly crescentic kidney partly embraces the heart and peri- cardium, just behind the origin of the penial retractor. Retractor muscles (Text-fig. 13 E) The cephalic retractor arises by two roots in the mid-dorsal line just behind the kidney, and divides about half-way along into right and left tentacular retractors. The right tentacular retractor passes forwards between the penis and oviduct, and not on the left of both as it does in Milax and Agriolimax. The pharyngeal retractor arises half-way along the left tentacular retractor, and divides into right and left branches passing to the buccal bulb between the cerebral and visceral ganglia. The penial retractor arises from the diaphragm on the left side, in front of the pericardium. Nervous system (Text-fig. 13 H) The olfactory and optic lobe of the cerebral ganglion is large, the cerebro-pleural and cerebro-pedal commissures short, and the visceral ganglia closely approximated, the right visceral and abdominal ganglia being fused together. Jaw (Text-fig. 13 F) Of the usual oxygnathous type, 1-5 mm across. Radula (Text-fig. 13 G) The central tooth, 0-038 X 0-018 mm, with small but distinct ectocones, is flanked on each side by about twenty lateral teeth with both ectocones and endocones, 1 82 BRITISH SLUGS followed by about forty-five marginal teeth. The teeth between the eighteenth and twenty-second are transitional in character, the endocone diminishing, and the ectocone approaching the mesocone. In the marginals, the mesocone becomes relatively longer and more curved, and the ectocone becomes prominent, arising from the side of the mesocone. Ectocones occur in some of the outer marginals of other Limax species, but they are not the conspicuous feature throughout that they are in L. tenellus. Formula .20.45 x no. FIG. 13. Limax tenellus. A. Reproductive organs (x 4). B. Penis (X 6-7) opened. G. Alimentary system, showing left lobe of liver forming apex of visceral mass, ovotestis at stomach level, between left and right lobes, intestine, cephalic retractor, oesophagus, salivary glands, buccal bulb, portion of albumen gland below left liver lobe. D. Kidney, heart and pericardium, origin of penial retractor in front, second forwardly-directed loop of intestine passing around cephalic retractor, rectum. E. Cephalic retractor muscles. F. Jaw ( x 14). G. Represent- ative radular teeth ( x 330). H. Nerve ganglia (x 8). BRITISH SLUGS 183 Alimentary system (Text-fig. 13 c) The left lobe of the liver forms the apex of the visceral mass. The apex of the stomach forms the most posterior part of the alimentary tract (which in Limax has almost lost the spiral torsion seen in Avion and in Milax), and the intestine has two forwardly directed loops. The first of these is held in place by the aorta as usual, and the second, which extends further forwards, is crossed by the cephalic retractor. The intestine runs backwards from this point and turns sharply to run forwards as the rectum, between the cephalic retractor and the floor of the pulmonary chamber, to the posterior lip of the respiratory orifice. Reproductive system (Text-fig. 13 A) The darkly pigmented ovotestis lies behind the stomach, but a little of it may be visible from the dorsal aspect, between the right and left lobes of the liver. The dilated central part of the hermaphrodite duct is pigmented and sinuous, narrowing at the albumen gland, where it has a distinct unpigmented vesicula seminalis. The prostate does not diverge anteriorly from the oviduct. The short vas deferens, which is not bound down by the right ocular retractor, becomes sacculate and pig- mented at the entrance to the penis, apically above the insertion of the penial retractor. Internally the penis (Text-fig. 13 B) shows three or four delicate short transverse folds apically, and further forwards three main massive longitudinal folds, the exact appearance varying according to where the penis is opened. These folds are visible during copulation (see below). The spermatheca is small and club- shaped, with a short duct entering the atrium. Mating According to Gerhardt (1933 : 445), one animal follows the other rapidly, head to tail. The leader soon bends to the right, so that after ten minutes the circle is completed. The penes are suddenly everted as massive bluish-white bodies, each with a thickened oblique fold which Gerhardt calls the comb. The penes are closely apposed, but not entwined, and the animals slowly revolve clockwise for one and a half hours. L. tenellus differs from L. maximus and L. cinereoniger in not entwining the body nor hanging from a horizontal surface. Development The eggs vary from 3-6 x 3-1 to 3-3 X 3-0 mm, and are pale amber, soft, trans- lucent and slightly adherent. In Britain, they are laid from November to March in clusters of about fifteen, and take from sixty to 120 days to hatch. The young are 6 mm long, with a slight dorsal keel posteriorly. They are white and translucent, without body and mantle bands and with a more opaque kidney region. The tentacles are pale violet, with the dark brown retractor muscle showing through. Ecology Restricted to woodlands, usually of considerable extent, and one of the few slugs found in pine woods. Absent from most of the many woods and plantations ZOOL. 6, 3. 13 184 BRITISH SLUGS established during the seventeeth century, which suggests that its powers of dispersal are poor. Indifferent to the presence or absence of lime. Distribution (Map 19) Of very local occurrence, though wide range, and therefore seldom encountered. In Britain, recorded from many vice-counties. In Europe, found from France northwards to Jemtland, and eastwards to the Caucasus. Material examined Living specimens from Netley Heath, Shere, Surrey ; and Chepstow, Monmouth- shire. Spirit specimens from Britain and Switzerland. Limax flavus Linnaeus, 1758 Limax flavus Linnaeus, 1758 : 652. Sweden. Limax variegatus Draparnaud, 1801 : 103. Limax megalodontes Quoy & Gaimard, 1824 : 426. First British record By Lister (1685-97, pi- ^-^ib). Diagnostic features Tentacles steely blue, body and mantle with yellowish spots and no lateral bands. Long rectal caecum and long cylindrical penis a diagnostic combination for British Limax. External appearance (PI. 2, fig. 22) A rather large slug, from 7 to 10 cm long when extended. Colour from yellowish to greenish, more or less suffused with dusky pigment, and mottled on the mantle with yellow patches, on the body with yellow spots formed by groups of up to seven tubercles. There are no lateral bands. The tentacles are a very characteristic cold steely blue, the sole pale yellow, the respiratory orifice pale-rimmed, and the mucus yellow. External variation Var. rufescens Moquin-Tandon (1855), reddish-brown ; var. virescens Ferussac (1819), uniformly greenish ; var. antiquorum Sowerby (1834 ?)> P a l e ochraceous marbled with grey ; var. breckworthiana Lehmann (1864), uniformly suffused with dark pigment. Shell Rather large, 9x6 mm, thin and convex, with the nucleus near the posterior margin and a little to the left of the mid-line. BRITISH SLUGS 185 Pallial organs As in L. tenellus. Retractor muscles (Text-fig. 14 D) The cephalic retractors arise from two roots, which do not unite for a considerable distance. The buccal retractor arises near the origin of the left ocular retractor. The penial retractor arises from the diaphragm, in the region of the kidney. A c B FIG. 14. Limax flavus. A. Alimentary system ( x 2) showing second forwardly- directed loop of intestine passing around cephalic retractor, long rectal diverticulum passing backwards. B. Reproductive organs (x 2). C. Penis (x 6) opened to show internal folds. D. Cephalic retractor muscles. E. Jaw (x 14). F. Representative radular teeth ( x 330). i86 BRITISH SLUGS Nervous system As in L. tenellus. Jaw (Text-fig. 14 E) 2-25 mm across, dark brown in colour. Marked transversely with fine striae, which are crossed by still finer ones following the contour of the jaw. Radula (Text-fig. 14 F) Measures 6-5 X 2-6 mm, with about 150 rows of teeth. The central tooth (measur- ing 0-05 X 0-038 mm) has a strong mesocone, but its ectocones are obsolete or absent. The lateral teeth are also devoid of ectocones and endocones. The aculeate marginals, some of the outermost of which have small ectocones, are connected to the laterals by several transitional teeth. Formula .16+4.50 X 150. Alimentary system (Text-fig. 14 A) The oesophagus dilates to a capacious crop, flanked by the salivary glands, and merging gradually into the stomach. The second forwardly directed loop of intestine is held in place by the cephalic retractor muscle. Instead of being prolonged back- wards as it is in L. tenellus, it soon turns forwards as the rectum ; but from the apex of the bend a long rectal diverticulum extends backwards, on the surface of the liver, to the end of the body. As in the other species of Limax, the left liver lobe forms the apex of the visceral mass. Reproductive system (Text-fig. 14 B) The unpigmented ovotestis lies between the lobes of the liver, well below the stomach apex. The hermaphrodite duct, at its entrance to the yellow linguiform albumen gland, has a rounded white seminal vesicle. The ribbon-like yellow prostate diverges from the oviduct anteriorly, and can easily be peeled off it for the greater part of its length. The vas deferens runs forward, and turns back beneath the right ocular retractor to the apex of the penis. The penis is a rather long cylindrical organ, strongly folded in a spiral. Internally it bears (besides three or four minor longitudinal folds) a long, low, crinkled, ribbon-like fold along its whole length, and another more prominent fold that expands distally to a crest or " comb ", lobulated or crenulated on the free margin (Text-fig. 14 c). The oviduct expands at its entrance to the atrium, where the spermatheca arises from it. Mating Mating occurs from late summer to February. According to Gerhardt (1933), after " following " for about eight minutes, the animals approximate the right sides, but do not curve into a circle. The penes are rapidly everted and entwine spirally, forming a mass of about one square centimetre in area. Sperms enclosed in a mucous mass are transferred between the penes, which are rapidly withdrawn carrying the partner's sperms with them. The whole process occupies only half BRITISH SLUGS 187 a minute. Gerhardt does not state what follows, but presumably the sperms are transferred from the withdrawn penes to the spermathecae. Development The eggs are laid in clusters of up to twenty, and have been found in September and February. They are large, 6x4 mm, pale amber in colour, and differ from those of other British slugs in having a small rounded projection at each end, so that they resemble miniature lemons. They hatch in from three to six weeks at room temperatures. The newly-hatched young are from 10 to 13 mm long, pale greenish-yellow with the characteristic blue tentacles already evident. There are no bands at any stage, and the yellow spots of the adult appear later in development. Behaviour L. flavus is essentially nocturnal, and unless looked for after dark it may escape observation in urban surroundings, even where it is abundant. Ecology L. flavus is found in woods, under logs and bark, but is much commoner in gardens, in the crevices of walls, and in cellars and out-houses. It eats decaying vegetable matter, fungi and lichens in wild places, and mildew and garbage in gardens and out-houses. Distribution (Map 20) Probably occurs in every vice-county of the British Isles, and ranges from southern Norway to North Africa and eastwards to Syria. It has been introduced into South Africa, Australia and North and South America, and to islands in the Atlantic and Pacific. Material examined Living specimens from Swansea, Glamorgan ; and Reading, Berkshire. Spirit specimens from Britain, Corsica, Greece, North and South America, South Africa, Australia, and the New Hebrides and Cook Islands. Limax cinereoniger Wolf, 1803 Limax cinereoniger Wolf, 1803 : 7. Germany. Limax antiquorum F6russac, 1819 : 68 (in part). Limax maximus', Gray, 1840 : 113 (in part). Avion lineatus Dumont, 1850 : 64. Limax cinereus var. intermedia Breviere, 1881 : 314. First British record By Gray (1840). i88 BRITISH SLUGS Diagnostic features Largest British slug, usually black with dorsal line and median zone of sole white, keel prominent and rather long. Distinguished from L. maximus by coloration, and by coarser tubercles, spotted tentacles, relatively small shell and jaw, and longer penis not tapering distally. External appearance (PI. 2, fig. 25) This is our largest native slug, which grows to 10 or 20 cm long in this country, while some continental forms are even larger. The keel is prominent and long, and the tubercles relatively coarser than in L. maximus. The slug is usually black all over, except for the white mid-dorsal line and median area of the sole. The dark tentacles are spotted with small discrete black or dark brown dots a character distinguishing the species from L. maximus, which has apparently not been recorded previously. External variation In Britain the ground colour, seen in the mid-line and the median area of the sole, is usually whitish, but in continental specimens the keel may be red or yellow. The following named varieties have been recorded from Britain : var. ornata Lessona (1880), body black with a series of white spots along each side ; var. maura Held (1836), entirely black ; var cinerea Moquin-Tandon (1855), entirely ash-coloured except for bluish-black mantle ; var. punctata Lessona (1880), keel and mid-line of dorsum yellow, two interrupted bands on each side. Shell Relatively smaller (9x5 mm), thinner and slightly more elongated than that of L. maximus. Pallial organs As in L. tenellus. Retractor muscles The slender cephalic retractor arises from the mid-line behind the diaphragm, and crosses the second forwardly-directed intestinal loop. The right ocular retractor passes (as is usual in Limax] between the male and female divisions of the distal genital ducts, over the vas deferens. The penial retractor arises broadly from immedi- ately to the left of the cephalic retractor, and narrows to its insertion at the apex of the penis. Nervous system As in L. tenellus. BRITISH SLUGS 189 Jaw (Text-fig. 15 E) Usually about 3 mm across, but may reach 4 mm in very large specimens : relatively smaller than in L. maximus. 23 1*0 bo bq 23 40 60 70 H FIG. 15. Limax. A. L. cinereoniger distal genital ducts ( x 1-3). B. L. maxi- mus mating pair suspended from mucus thread (adapted from L. Adams). C. L. maximus distal genital ducts (x 2). D. L. maximus stages in eversion and entwining of penes (adapted from L. Adams). E. L. cinereoniger jaw (x 6-7). F. L. maximus jaw (x 6-7). G. L. cinereoniger representative radular teeth ( x 33) H. L. maximus representative radular teeth ( x 330). igo BRITISH SLUGS Radula (Text-fig. 15 G) There are about 150 rows of teeth. The central tooth has rather small but strong ectocones. It is flanked by about nineteen laterals, with ectocones and endocones, the latter descending on the mesocones in the outer teeth. There are about fifty marginals on each side, of which only the inner teeth retain the endocone. In the outer marginals, an ectocone appears on the concave side of the mesocone, and becomes bifid or trifid in a few of the outermost teeth. Formula .19.50 x 150. Alimentary system Resembles L. tenellus, and differs from L. flavus, in having the loop of intestine which is retained by the cephalic retractor prolonged backwards almost to the visceral apex, and without a rectal caecum. The loop is longer than in L. maximus. Reproductive system (Text-fig. 15 A) The long narrow darkly pigmented ovotestis is visible on the surface between the left and right liver lobes. The hermaphrodite duct is pale, slender at first, then wider and folded, and narrows at the somewhat triangular albumen gland. If a seminal vesicle is present, it is small, deep in the albumen gland and difficult to find. The folded common duct narrows to a long free oviduct, dilated at its entry into the atrium, and the small club-shaped spermatheca arises one or two millimetres before its end. The prostate separates from the common duct and so is free at its anterior end, and the vas deferens passes under the right ocular retractor and back to the apex of the penis. The stout cylindrical contorted penis is of nearly uniform diameter throughout its length, and when straightened out is from 6-8 cm long. Internally it bears a prominent fold, double at its origin. In its upper part this expands into a prominent frill, the " comb ". The rest of the mucous membrane is circularly ridged. Mating Mating occurs at night, throughout the milder periods of the year, on vertical or overhanging surfaces. After about fifteen minutes of " following ", the animals hang down freely from a disk of mucus. Their bodies and the evaginated penes, which become as long as the bodies, are entwined. A mass of sperms enclosed in mucus descends each vas deferens inside the penis, and the sperm packets are exchanged between the "combs" which are now terminal. The penes are then withdrawn, carrying the sperms with them to be transferred later to the spermathecae. The whole process takes about twenty minutes. Auto-fecundation can occur in this species (Oldham, 1942^). Development The eggs are about 5 X 5-5 mm, soft, translucent and amber-coloured. On account of the large size of the eggs, and their transparency when under water, the caudal and cephalic vesicles of the embryos can easily be observed (Text-fig. 18 H). The BRITISH SLUGS 191 eggs hatch in about a month. The newly hatched young are translucent white, but soon become opaque, then brown and finally black. There is sometimes a sug- gestion of banding, and the lateral areas of the foot are the last to pigment. Oldham (1942^) records that this slug becomes mature in two years, and that an individual in captivity lived for five years and 270 days. Ecology This is essentially a slug of wild places, being found beneath logs and leaves and in tree crevices of ancient woodlands. In Pembrokeshire at least, it also occurs on damp lichen-covered rocks on bare hillsides. It seems never to occur in gardens and cultivated land or recent plantations. Though in the wild it presumably feeds on lichens, fungi and decaying vegetable matter, in captivity it readily eats carrot, oatmeal etc. Distribution (Map 21) L. cinereoniger is recorded from most British vice-counties, except for East Anglia and parts of central England, southern Scotland and central Ireland. In Europe, it extends from arctic Norway to north-eastern Spain, Italy, Sardinia, Greece and the Crimea. Material examined Living specimens from Netley Heath, Surrey ; Cusop Dingle, Hereford ; Buckle- bury, Berkshire ; and Pont Neath Vaughan, Glamorgan. Spirit material from Britain, Denmark, Germany and Luxemburg. Limax maximus Linnaeus, 1758 Limax maximus Linnaeus, 1758 : 652. Sweden. Limax cinereus Miiller, 1774 : 5. Limax antiquorum F6russac, 1819 : 68 (in part). First British record By Lister (1674 : 99). Diagnostic features Large, differs from L. cinereoniger in coloration : grey or brown with two or three bands on each side of body, dark spotting or marbling on mantle, tentacles pale pinkish-brown without spots, sole uniformly pale ; and also in having the tubercles relatively smaller, the shell and jaw relatively larger and the penis shorter and tapering distally. External appearance (PI. 2, fig. 24) Length from 10 to 20 cm. The tubercles are relatively smaller than in L. cinereoniger. The slug is usually some shade of grey, with a uniformly pale sole. IQ2 BRITISH SLUGS The body bears two or three bands, often doubled or more or less interrupted, on each side. The mantle is irregularly spotted or marbled with dark pigment, but never banded. The tentacles are translucent pinkish-brown, without spots. External variation The ground colour may be grey, ochraceous or reddish, and the bands may be developed to an infinitely variable degree. Named varieties are : var. concolor Pini (1876), uniformly ash-coloured ; var. Candida Lessona & Pollonera (1882), white, translucent, eyes pale brown ; var. vinosa Baudon (1884), purple or vinous brown ; var. fasciata Razoumowsky (1789), ash-coloured with three bands on each side ; var. sylvatica Morelet (1845), ash-coloured with two bands, and an indistinct third lower band ; var. tetrazona Taylor (1902), with only two bands on each side ; var. cellaria Dezallier d'Argenville (1742), bands interrupted ; var. aldrovandi Moquin-Tandon (1855), ash-coloured with pale spots ; var. tigris Taylor (1902, from Adams MS.), tawny yellow with black bands. Shell Measures about 11x7 mm, and is thus relatively larger than that of L. cinere- oniger. Pallial organs, retractor muscles, nervous system As in L. cinereoniger. Jaw (Text-fig. 15 F) Relatively larger (3-5 mm across) than that of L. cinereoniger. Radula (Text-fig. 15 H) The central tooth is larger and relatively narrower than in L. cinereoniger, and its ectocones are more delicate and sometimes almost obsolete. There are about nineteen lateral teeth on each side, all or most of which retain both ectocones and endocones. The marginals (about fifty on each side) lack endocones, and only some of the outermost teeth have small ectocones. Formula 0.19.50. Alimentary system As in L. cinereoniger, except that the last posteriorly-directed loop of intestine is shorter, reaching only about half-way along the visceral mass. There is no rectal caecum. Reproductive system (Text-fig. 15 c) The ovotestis is of the same long flattened linguiform shape as in L. cinereoniger, but is pale or only slightly pigmented. The hermaphrodite duct, albumen gland, common duct, free oviduct and spermatheca are also similar. The distal end of the prostate diverges more or less from the oviduct, and above this the prostate is BRITISH SLUGS 193 only weakly attached for a considerable distance. The penis differs from that of L. cinereoniger in being shorter, widest apically and tapering distally. Internally there is a fold, expanded proximally into a " comb ", as in L. cinereoniger. Mating (Text-figs. 15 B and D) This occurs at night from overhanging surfaces. Unlike L, cinereoniger, L. maximus hangs suspended in mid-air from a stout thread of mucus. The protrusion of the organs, sperm transfer, and withdrawal of the penes is as in L. cinereoniger, but the entwined penes are not so long. After withdrawal, the animals re-ascend the mucus thread. Adams (1898) gives further details. Development The eggs, 5-0 x 5-5 mm, resemble those of L. cinereoniger. They are laid in early spring and in the autumn. February eggs in an unheated room hatch in six and a half weeks. The young are pale grey, with the middle lateral bands visible from hatching thus differing from young Lehmannia marginata, in which the upper band appears first and mantle bands are also present. The tentacles are pale pinkish- grey. After two or three weeks, the other body bands have appeared and begin breaking up, and the mantle becomes marbled. L. maximus lives for at least three or four years. Ecology L. maximus occurs in woods, hedgerows and waste places, where it is sometimes found together with L. cinereoniger ; but unlike the latter it occurs also in gardens, cellars and out-houses. It seems to avoid green leaves and to feed on fungi and decaying matter, though in captivity it readily eats oatmeal, Bemax etc. Distribution (Map 22) Apparently general in the British Isles, and found from Oslo and Petrograd to Asia Minor, Algeria and the Atlantic isles. Introduced into North America, South Africa, Australia, Tasmania and New Zealand. Material examined Living specimens from Swansea, Glamorgan ; Savernake Forest, Wiltshire ; Bucklebury Common, Berkshire. Spirit material from Britain, France, Greece, Azores, Madeira, Canary Isles and British Columbia. Genus LEHMANNIA Heynemann, 1861 Type species Limax marginatus Miiller, 1774. The penis is short, with a laterally-placed apical appendix (conical in L. marginata, short and cylindrical in L. poirieri, and long and cylindrical in L. melitensis Lessona & Pollonera). A long rectal caecum extends to the apex of the visceral mass. In 194 BRITISH SLUGS L. poirieri and L. melitensis the radula is like that of Agriolimax ; but that of L. marginata is highly specialized. The body readily absorbs water and becomes swollen and translucent. Hesse (1926) adopts sectional names within Lehmannia, segregating L. marginata (Lehmannia s.s.) from L. poirieri and L. melitensis (Ambigolimax] because of the former's specialized radula, although their reproductive and alimentary systems are very similar. As for Limax, subdivision of the genus depends on the importance ascribed to different characters. Lehmannia marginata (Miiller, 1774) Limax marginatus Miiller, 1774 : 10. Denmark. Limax arborum Bouchard-Chan tereaux, 1837 : 164. First British record By the Reverend B. J. Clarke (in Thompson, 1840 : 204). Diagnostic features Gelatinous aspect in damp weather, mantle bands forming lyre-shaped figure, two body bands on each side ; conical penial appendix, rectal caecum, specialized radula. External appearance (PI. 2, figs. 18 and 21) Usually about 7-5 cm long, or larger. Its great capacity for absorbing water makes this slug become semi-transparent and gelatinous in appearance during damp weather. When irritated, it exudes copious watery mucus. In colour it is character- istically grey with a pale sole, two darker body bands on each side (which may be more or less indistinct or interrupted), and a pair of mantle bands forming an elongated lyre-shaped figure. The keel is paler than the body, and the median area of the mantle often darker. External variation Var. glauca Clarke (1843), ground colour greenish-grey ; var. subrufa Le Compte (1871), ground colour yellow ; var. rosea Broeck (1870), ground colour pinkish ; var. bettonii Sordelli (1870), body bands fused together and interrupted; var. tigrina Wienland (1876), with rows of black spots on body and mantle ; var. rupicola Lessona & Pollonera (1882), darkly pigmented, with indistinct markings. Shell White and iridescent. Normally measures 4 X 2-75 mm, but varies considerably in size and thickness. Pallial organs As in Limax, BRITISH SLUGS 195 Retractor muscles (Text-fig. 16 c) The main stem of the cephalic retractors arises from the mid-dorsal line just behind the kidney, and divides about half-way into the right and left ocular retractors. The buccal retractor arises from the left ocular retractor, a short distance in front of the bifurcation. The strong penial retractor arises from the diaphragm between the pericardium and the posterior border of the kidney, to be inserted below the vas deferens. 15 50 51 10 FIG. 1 6. Lehmannia marginata. ( X 5-3) opened longitudinally, tract and salivary glands. E. (X 330). A. Reproductive organs (x 3-3). B. Penis G. Cephalic retractor muscles. D. Alimentary Jaw ( x 13). F. Representative radular teeth ig6 BRITISH SLUGS Nervous system As in Limax. Jaw (Text-fig. 7 E) 1-4 mm across, pale yellow in colour, and of the usual Limacid shape, not as narrow as figured by Taylor (1902-07, fig. 106). Radula (Text-fig. 16 F) Very characteristic, and easily distinguishable from the radulae of other British slugs. The central tooth has a strong, broad and short mesocone, and very small or obsolescent ectocones. It is flanked by ten or eleven laterals, with short wide mesocones and no ectocones, of which only the innermost one or two may have small endocones. The numerous marginals (up to seventy on each side) have charac- teristically strong, arched reflections bearing the short wide mesocones. From about the fortieth marginal outwards, the mesocone becomes bifid and then trifid, and minute denticulations corresponding to ectocones appear on the reflection. Alimentary system (Text-fig. 16 D) Resembles that of Limax flavus, with a rectal caecum extending backwards on the surface of the liver to the level of the stomach apex. The visceral cavity is darkly pigmented. Reproductive system (Text-fig. 16 A) The ovotestis is bulky and compact and superficially pigmented. The hermaphro- dite duct is pale, swollen and folded in its median portion, and bears a rounded seminal vesicle where it terminates at the short linguiform albumen gland. The oviduct and spermatheca resemble those of Limax. The prostate is extraordinarily massive and deeply pigmented. The vas deferens, bound down by the right ocular retractor muscle, is short and wide, and enters the penis at the side of its blunt apex. At the opposite side of the apex there is a short conical appendix. The penis opened longitudinally (Text-fig. 16 B) shows a prominent fold, expanded and free at the apex and almost smooth, but corresponding with the " comb " of Limax. There is a small conical fold at the entrance to the appendix. The walls of the proximal part of the penis are thick and rugose. Mating This species can mate on level, vertical or overhanging surfaces. Development In Britain, mating occurs during the winter months. December eggs in an unheated room hatch in February. The eggs are soft, very translucent and pale amber in colour. They usually measure 4-3 X 3-6 mm, but some are smaller. The young are 10 mm long, pale translucent grey with violet tentacles. They show well-marked BRITISH SLUGS 197 lyriform mantle bands, and a distinct band high up on the body in contrast to young Limax maximus, in which the middle band is the first to appear. Behaviour Like many (perhaps most) slugs, especially when young, L. marginata is able to suspend itself from a mucus thread secreted by the pedal gland. This thread passes back along the foot to the tip of the tail and adheres to a branch or other object, and supports the animal when it crawls off. Sometimes the slug can even turn around and re-ascend the thread. Kew (1902) gives numerous examples of this process, with details and some figures. Ecology Not uncommon in old and newer woodlands, especially in the west and north. In dry weather it retreats into crevices in the bark and between roots, while in damp weather it crawls up the trunks to a height of twelve feet or more. It is also found on damp rocks on exposed hillsides, and on old stone walls. Distribution (Map 23) Probably occurs in every vice-county of the British Isles, but is commoner in the west and north. In Europe it is found from Iceland and Lapland to Russia and Italy, and is recorded as introduced into Australia and New Zealand. In North America, it apparently occurs wild at Torbay, Newfoundland (confirmed by Altena, 1950). However, introduced slugs from Verdugo Woodlands, Glendale, California, kindly sent by Dr. W. O. Gregg and recorded as this species, proved to be L. poirieri. Probably some of the records from southern Europe also refer to L. poirieri. Material examined Living specimens from the Gower Peninsula and the neighbourhood of Swansea, Glamorgan ; Brecknock ; Caernarvon ; and Torrington, Devon. Spirit material from Britain and Australia. GREENHOUSE ALIENS Lehmannia poirieri (Mabille, 1883) ? Limax valentianus Ferussac, 1823, 2 : g6 e . Limax poirieri Mabille, 1883 : 52. Spain. Notes A full list of references is given by Altena (1950 : 9-17). First British record By Quick (1949 : 24), as L. valentianus. However, Dr. A. D. J. Meeuse had already found it in a greenhouse in Belfast Botanical Gardens, on 26th November, 1948. ig8 BRITISH SLUGS Diagnostic features Somewhat like L. marginata, but the ground colour is yellower and paler, and the body bands nearer to the mid-dorsal line. The radula is like that of Agriolimax, very different from the specialized radula of L. marginata. The penial appendix is longer and more cylindrical than that of L. marginata, and not pointed at the apex. External appearance (PL i, fig. 14) L. poirieri is about 6 cm long, with a yellowish-grey body and mantle, and slightly darker head. The general appearance is watery and translucent, like a pale L. marginata. The keel is short and inconspicuous. A pair of body bands, high up near the mid-dorsal line, are usually more or less interrupted and with scalloped outer edges. Sometimes faint traces of a second pair of bands occur lower down. A pair of mantle bands, becoming broader and darker behind, form a lyre-shaped figure, and the area enclosed is often darker in the centre and mottled with brown. The respiratory orifice is pale bordered. The tripartite sole is uniformly pale. External variation The mantle and body bands vary in width, completeness and the degree of crenu- lation of their outer borders, and there are sometimes faint traces of a second pair of body bands. Jaw Pale yellow, 2-25 mm across. Ra&ula (Text-fig. 17 D) Like that of Agriolimax in having a distinctly tricuspid central tooth, and endo- cones on the lateral teeth a feature which led Pollonera (18870 : 2) to place the species erroneously in that genus. The tri-cuspid central tooth is flanked by fifty-five to sixty teeth on each side, of which about fifteen are typical laterals with endo- and ectocones. Most of the inner marginals are simply aculeate, but many of the outer ones are bifid, while in the extreme outer ones the main cusp becomes multi- denticulate. Alimentary system The intestine shows the two forwardly-directed loops and the long rectal caecum reaching to the apex of the visceral mass, as in L. marginata and Limax flavus. Reproductive system (Text-fig. 17 A) Similar to that of L. marginata, but with the penial appendix longer, and bluntly cylindrical rather than conical. Pollonera (18870) figures the appendix of valentianus as dilated and terminally vesicular, but this may result from misinterpretation of a somewhat folded appendix (Simroth, 1887). Altena (1950) found that Canadian BRITISH SLUGS 199 specimens had simple cylindrical appendices, like those of specimens from Swansea. Another difference from L. marginata is that the sac into which the spermatheca duct expands is somewhat sausage-shaped, rather than pyriform. 15 25 35 JO FIG. 17. Greenhouse aliens. A. Lehmannia poirieri genital organs (x 3-3). B. Limax nyctelius genital organs (x 3-3). G. L. nyctelius penis (x 6-7) opened longitudinally. D. L. poirieri representative radular teeth (x 330). E. L. nyctelius representative radular teeth (x 330). ZOOL. 6, 3. 200 BRITISH SLUGS Mating, development No observations on mating seem to have been recorded. Eggs laid in captivity at Swansea in December 1939 were translucent yellow and measured 2-25 X 1-5 mm. They hatched in February. In March the young measured 10-0 mm long, and had intensely black body and mantle bands, and by loth June they were full grown. Ecology Found always on the ground, never climbing trees and plants like L. marginata. This is true in Tenerife (Altena, 1950) and California (Gregg, in litt.}, and in a green- house at Swansea, where the species was always found under pots. Distribution Spain; the Canary Isles (? introduced) , and introduced into California and Britain (known from greenhouses in Reading, Swansea and Belfast). Material examined Living specimens taken from a greenhouse in Singleton Park, Swansea, Glamorgan, in 1936, 1938 and 1949 ; and from a Reading University greenhouse at Shinfield, Reading, in 1959. Spirit specimens from Verdugo Woodlands, Glendale, California (collected by Dr. W. O. Gregg) ; Guejar de la Sierra, Provincia de Granada, South Spain, and Elizondon, Navarre, North Spain (collected by Dr. de Zarate) and Barcelona. Limax nyctelius Bourguignat, 1861 Limax nyctelius Bourguignat, 1861 : 305, pi. 2, figs. 3 and 4: Algeria. First British record By Quick (1949 : 25). However, A. R. Waterston had found it some years before this, and partially described the species in an unpublished thesis. Diagnostic features Externally like Lehmanniapoirieri, and with a similar radula. Internally resembles L. flavus in having a long rectal caecum, and a long cylindrical penis without an appendix. External appearance L. nyctelius is about 5 cm long, pale greyish-yellow with a thin translucent body- wall. There is a short inconspicuous keel at the hinder end. The mantle bands form a lyre-shaped figure, and the narrow body bands are high up on each side near the mid-dorsal line. The respiratory orifice has a pale border, and the sole is uniformly pale. BRITISH SLUGS 201 External variation The mantle and body bands vary somewhat in breadth and intensity. Like those of Lehmannia poirieri, they tend to fade in alcohol. Jaw Dark brown or blackish, 1-7 mm across. Radula (Text-fig. 17 E) A tri-cuspid median tooth is flanked on each side by thirteen to fifteen tri-cuspid laterals and from twenty-eight to thirty aculeate marginals. The marginals beyond about the nineteenth are bifid, and the extreme marginals trifid. Alimentary system As in L.flavus, with two forwardly-directed loops and a long rectal caecum. Reproductive system (Text-fig. 17 B) The pigmented ovotestis is exposed between the right and left lobes of the liver. The pigmented hermaphrodite duct becomes swollen and folded in its central portion, and ends at the large albumen gland with a pale oval seminal vesicle. The free oviduct is short and stout. The spermatheca arises by a wide duct from the atrium, which expands abruptly to an oval sac which may be pigmented. The penis is uniformly cylindrical, more or less folded, and the vas deferens enters its apex under cover of the stout penial retractor muscle, which arises from the diaphragm close behind the kidney. Internally the penis (Text-fig. 17 c) resembles that of L. flavus in containing a prominent fold starting at the apex and expanding below to a " comb ", and another shorter smooth fold. Ecology Recorded from greenhouses and garden frames in Britain, on date palms in Washington, D.C., and beneath flower pots in Egypt (Abu Teira, in litt.}. Distribution North African coast and Egypt. Introduced into Britain (known from greenhouses in Edinburgh and Glasgow), North America (known from Washington, D.C.) and South Africa (Connolly 1939 : 176). Material examined Spirit specimens from the Royal Botanic Gardens, Edinburgh (sent by A. R. Waterston), Shebin El Kom, Egypt, Algeria and Washington, D.C. 202 BRITISH SLUGS FOSSIL SPECIES Limax modioliformis Sandberger, 1880 Limax modioliformis Sandberger, 1880 : 103, pi. xii, figs. 15-15 c. Fossil shells ascribed to this species are found in the Cromerian at West Runton, Norfolk (Sandberger, 1880). However, the specific ascription of fossil slug shells is at best only tentative. Similar shells in the same beds are ascribed to Lehmannia marginata. FIG. 18. Embryonic vesicles. A. Achatina fulica embryo (x 6-7) showing large caudal vesicle. B. A. fulica later embryo (x 6-7) with vesicle adsorbed. C. TestacMa scutulum embryo (x 6-7) showing cephalic and caudal vesicles. D. Helix aspevsa embryo (x 13) showing caudal vesicle, cephalic vesicle adsorbed. E. Helice-lla itala embryo ( x 20) showing caudal vesicle, cephalic vesicle adsorbed. F. Avion ater embryo (x 17) three weeks after laying, with cephalic and caudal vesicles. G. Agriolimax caruanae embryo ( x 17) with cephalic and caudal vesicles. H. Limax cinereoniger embryo ( x 6-7) with cephalic and caudal vesicles. BRITISH SLUGS 203 ANATOMICAL TERMS The following terms are used throughout this paper, in discussing the genital anatomy of slugs. They are illustrated by somewhat generalized illustrations in Text-fig. 19. The hermaphrodite gonad, far back in the visceral mass, producing ova and spermatozoa. The slender duct conveying ova and spermatozoa from the ovotestis to the albumen gland, oviduct and prostate. A loop, enlargement or diverticulum of the hermaphrodite duct at the albumen gland, for the storage of the animal's spermatozoa. A more or less linguiform gland at the termination of the herma- phrodite duct, supplying the egg albumen. A cavity at the commencement of the oviduct where the ova are fertilized by the partner's spermatozoa. Glandular tubules, opening into the male duct or groove which conveys the spermatozoa. The combined prostate, male duct or groove and oviduct. The prostate usually appears as a yellow ribbon on the gelatinous- looking oviduct. In the Stylommatophora, the proximal parts of the male and female tracts are not completely separate from one another. The slender male duct after it becomes separate and leaves the sperm-oviduct. The female duct, beyond the point where the vas deferens leaves it. The expanded termination of the spermatheca duct, which arises near the termination of the free oviduct (Milax) or directly from the atrium (Arion, Limax and Agriolimax). The spermatheca temporarily stores the partner's spermatozoa. The part of the free oviduct between the origin of the spermatheca duct and the atrium. When the spermatheca duct arises directly from the atrium, there is no vagina. The terminal portion of the genital ducts, receiving the oviduct, penis and often the spermatheca duct. The enlarged terminal portion of the vas deferens, which secretes the spermatophore and leads to the penis. A chitinous capsule containing a mass of spermatozoa (Arion, Milax). The terminal evaginable portion of the male duct, leading to the atrium, and sometimes (Milax) containing a penial papilla. A muscle arising from the body-wall and inserted between the epiphallus (when one is present) and the proximal end or apex of the penis. In Arion a penis and retractor are not present ; the epiphallus enters the atrium direct, and a retractor muscle is inserted on the free oviduct and the spermathecal duct. A slender tubular diverticulum from the apex of the penis (Testacella). A conical or digitiform appendage of the penis (Lehmannia), or a more or less branched appendage at the apex of the penis (Agriolimax). A stimulatory organ situated in the penis (Agriolimax). A stimulatory organ situated in the atrium (Milax) . A lobulated or frilled mass in the atrium or lower part of the oviduct in Arion, which in copula functions as an adhesive organ between the partners. A mass of glandular tubules in Milax, discharging into the termina- tion of the oviduct and the atrium. Ovotestis Hermaphrodite duct Seminal vesicle A Ibumen gland Fertilization sac Prostate Sperm-oviduct Vas deferens Free oviduct Spermatheca Vagina A trium Epiphallus Spermatophore Penis . Retractor penis muscle Flagellum Penial appendix Sarcobelum Stimulator Ligula Vaginal glands BRITISH SLUGS A.Testacella s p UA S P v D FO C. Mi lax vc VD D* Agriolimax S B H D OT S PA FIG. 19. Genital anatomy. A. Testacella. B. Avion. C. Milax. D. Agriolimax. Key : A atrium AG albumen gland E epiphallus F flagellum FO free oviduct FS -fertilization sac HD hermaphrodite duct L ligula LA lower atrium OT ovotestis P penis PA penial appendix PR prostate R retractor muscle SB sarcobelum SP spermatheca SPD spermatheca duct SPO spermoviduct ST stimulator SV seminal vesicle UA upper atrium V vagina VD vas deferens VG vaginal glands. BRITISH SLUGS 205 99SZJSM'Z M^O O O O tww )\O txOO 66 Durham 67 Northumb. S. 68 Cheviotland 69 Westmorland 70 Cumberland 71 Isle of Man % z g a a w s . 3 -g 1 < - g^^^a *& gw -^ oliis ilimtlnv'j2 | i ^ Jl S"0 ^3J*J-g3Sg33.fM ft 1 aS-S'eM'8. Ha S ^ ^SSliS J w 'llllill "i-iillii w "illi 2 M co T^ in'O tN. ooc\OHWfOTj- ino rxoo o\ O H N en TJ- m^o 206 BRITISH SLUGS - pq h .3 (S BRITISH SLUGS 207 S' .S I d 208 BRITISH SLUGS ffl d PQ d _o 4^ a ft c BRITISH SLUGS 209 f 210 BRITISH SLUGS BRITISH SLUGS BRITISH SLUGS 1 5 BRITISH SLUGS 213 2I 4 BRITISH SLUGS BRITISH SLUGS 215 ^=>s! ""# ZOOL. 6, 3. 2l6 BRITISH SLUGS a 'rt -t-> 'iH cc d BRITISH SLUGS REFERENCES ADAMS, L. E. 1898. Observations on the pairing of Limax maximus L. /. Conch, 9 (3) : 92-95 (pi. iii). - 1910. Observations on the pairing of Arion ater (L.). Ibid. 13 (4) : 116-119 (6 text-figs). 1941. Testacella scutulum method of cleansing. Ibid. 21 (8) : 223. ALDER, J. 1848. A Catalogue of the mollusca of Northumberland and Durham. Trans. Tyneside Naturalists' Field Club 1 (2) : 97-209. ALLMAN, G. J. 1843. On a new genus of terrestrial gastropod. Athenaeum (Sept.) : 851, col. 3. 1846. Description of a new genus of pulmonary gasteropods. Ann. Mag. nat. Hist. 17 : 297-299 (pi. xvii, A, figs. 1-3). ALTENA, C. O. VAN R. 1949. List of localities visited during a collecting trip to the Canary Islands in the spring of 1947. Tijdschr. Ent. 91 : 1-5. 1950. The Limacidae of the Canary Islands. Zool. Verh. Leiden (n) : 3-34. 1958. Notessur Ieslimaces5. Les Philomycidae et Limacidae de la republique El Salvador. Arch. Molluskenk. 87 : 27-31 (2 figs.) ARGENVILLE, A. J. DEZALLIER D'. See Dezallier d'Argenville, 1742. BABOR, J. F. 1894. Uber den Cyclus der Geschlechtsentwicklung der Stylommatophoren. Verh. dtsch. zool. Ges. (4) : 55-61 (10 text-figs). BARNES, J. F. & STOKES, B. M. 1951. Marking and breeding Testacella slugs. Ann. appl. Biol. 38 (2) : 540-545. BARNES, J. F. & WEIL, J. W. 1945. Slugs in gardens : their numbers, activities and distribu- tion. Part 2. /. Anim. Ecol. 14 (2) : 71-105 (pis. 3-5). BAUDON, A. 1884. Troisieme catalogue des mollusques vivants du departement de 1'Oise. /. Conchyliol. 32 : 193-325. BINNEY, A. 1843. Descriptions of some of the species of naked, air-breathing mollusca, inhabiting the United States. Boston Journ. nat. Hist. 4 (2) : 163-175. 1844. On some of the species of naked pneumonobranchus mollusca of the United States. Proc. Boston Soc. nat. Hist. 1 : 51-52. BINNEY, W. G. & BLAND, T. 1869. Land and freshwater shells of North America. Pt. i. Pulmonata Geophila. Smithson. misc. Coll. 8 (3) : i-xii + i-3!6 (545 figs.). BLAND, T. See BINNEY & BLAND, 1869. BONELLI, . In Lessona and Pollonera, 1882. BOUCHARD-CHANTEREAUX, N. R. 1837. Catalogue des mollusques terrestres et fluviatile observes jusqu'a ce jour a 1'etat vivant dans le departement du Pas-de-Calais. Mem. Soc. Agric. Boulogne (n) : 141-230. BOUILLET, J. B. 1835. Catalogue des especes et varietes de mollusques terrestres et fluviatiles observes jusqu'a ce jour a etat vivant, dans la Haute et la Basse- Auvergne, etc. Ann. Scient. Acad. Sci. Clermont-Ferrand 8 : 521-594. BOURGUIGNAT, J. F. i86i. Des Limaces Algeriennes. Rev. Mag. Zool. (2) 13 : 299-306. (Re- printed in Spicileges Malac. pp. 33-43 with new continuation, 2 pp.) BOYCOTT, A. E. 1934. The habitats of land mollusca in Britain. /. Ecol. 22 (i) : 1-38 (i fig.). BREVIERE, L. 1881. Tableau des limaciens des environs de Saint-Saulge (Nievre). /. Con- chyliol. 29 : 306-316. BROECK, E. V. 1870. Excursions decouvertes et observations malacologiques faites en Belgique pendant 1'anee 1870. Ann. Soc. zool. malac. Belg. 5 : 13-64 (pi. 2 ; i text-fig.). CAIN, A. J. & WILLIAMSON, M. H. 1958. Variation and specific limits in the Arion ater aggre- gate. Proc. malac. Soc. Lond. 33 (2) : 72-86 (5 figs.). CHEMIN, E. 1939(2. Observations sur la Testacelle : Testacella europaea Roissy = T. haliotidea Drap. Trav. Sta. zool. Wimereux 13 : 77-85 (pi. 2 ; i fig.). 19396. L'autofecondation chez Testacella europaea Roissy (= T. haliotidea Drap.). C.R. Soc. Biol. Paris 132 (24) : 249-252. ZOOL. 6, 3. i5 2i8 BRITISH SLUGS CLARKE, B. J. 1843. On the species of the genus Limax occurring in Ireland. Ann. Mag. nal. Hist. 12 : 332-342 (3 pis.). In Thompson, 1840. COCKERELL, T. D. A. 1885. The variation and abnormal development of the mollusca. Part 3. Terrestrial Gasteropoda. Hardwicke's Science Gossip 21 (250) : 224-226 (5 figs.). i886a. Notes. The Garner and Science Recorder's Journal 1 (9) : 139. 18866. Notes on some varieties of British shells. /. Conch. 5 (3) : 79-80. i886c. Some varieties of British shells. The Naturalist's World 3 (33) : 179. 18900. The British naturalist's catalogue of the land and freshwater mollusca of the British Islands. The Young Naturalist 11 (127) : 19 pp. 18906. Notes on slugs, chiefly in the collection of the British Museum. Ann. Mag. nat. Hist. (6) 6 : 277-288. 1891. Notes on slugs, chiefly in the collection of the British Museum. Ibid. (6) 7 : 328-341. 1892. A revised list of the species of British Slugs. /. Conch. 7 (3) : 66-70. COLLINGE, W. E. 18920. Descriptions of a new variety of Arion hortensis Fer. and A. circum- scriptus Johnst. The Conchologist 2 (2) : 26-27. 18926. A review of the Arionidae of the British Isles. Ibid. 2 (3) : 56-66. i892c. A review of the Arionidae of the British Isles. Ibid. 2 (4) : 76-83. 1893. On the occurrence of Arion lusitanicus Mab. in the British Isles, and descriptions of four new varieties. Ann. Mag. nat. Hist. (6) 12 : 414-415. 1894. The anatomy and description of a new species of Arion. Ibid. (6) 13 : 66-67 (pi- V.A.). 18950. Amalia parryi, a supposed new species. Journ. Malac. 4 (i) : 7. 18956. Notes on some slugs from Algiers. Proc. malac. Soc. Lond. 1 (7) : 336-337 (pi. xxiii) . 1896. On a collection of slugs from the Sandwich Islands. Ibid. 2 (i) : 46-51 (6 figs.). 18970. Some observations on certain species of Arion. Journ. Malac. 6 (i) : 7-10 (pi. 2). 18976. On some European slugs of the genus Arion. Proc. zool. Soc. Lond. (3) : 439-450 (pis. xxix-xxxi). iSgjc. On a further collection of slugs from the Hawaiian (or Sandwich) Islands. Proc. malac. Soc. Lond. 2 (6) : 293-297 (9 figs.). 1898. Note on a new variety of Testacella maugei Fer. /. Conch. 9 (3) : 95. COMTE, T. LE. 1871. Deuxieme liste supplemental des mollusques terrestres et fluviatiles recueilles aux environs de Lessines. Butt. Soc. mal. Belg. 6 : 6568. CONCHOLOGICAL SOCIETY CENSUS. See Ellis, 1951. CONNOLLY, M. 1939. A monographic survey of South African non-marine mollusca. Ann. S. Afr. Mus. 33 (i) : i-iii + 1-660 (pis. I-XIX ; 58 text-figs.). COOPER, J. G. 1863. On a new genus of terrestrial mollusca inhabiting California. Proc. Calif. Acad. Sci. 3 (i) : 62-63 (* n g-)- DEZALIER D'ARGENVILLE, A. J. 1742. L'Histoire Naturelle eclaircie dans deux de ses parties principales, la lithologie et la conchyliologie . . . viii + 491 pp. (33 pis.) 4to. Paris. D'ORBIGNY, A. D. See Orbigny, A. D. d', 1836-44. DRAPARNAUD, J. P. R. 1801. Tableau des mollusques terrestres et fluviatiles de la France. 116 pp. 8vo. Montpellier & Paris. 1805. Histoire naturelle des mollusques terrestres et fluviatiles de la France, i-viii + 1-164 pp. (13 pis. i port.) 4to. Paris. DROUET, H. 1868. Mollusques terrestres et fluviatiles de la Cote-d'Or. Mem. Acad. Dijon. (2) 14 : 33-154. DRUCE, G. C. 1932. The comital flora of the British Isles, xxxii -f 47 PP- Arbroath. DUMONT, F. 1850. Description d'une espece nouvelle du genre Arion (A. lineatus). Bull. Soc. Hist. nat. Savoie 1 : 64-65. DUMONT, F. & MORTILLET, G. 1856. Catalogue critique et malacostatique des mollusques de Savoie et du bassin du Leman. Bull. Inst. nat. Genevois 4 : 310-361. BRITISH SLUGS 219 ELLIS, A. E. 1926. British Snails. A guide to the non-marine gastropoda of Great Britain and Ireland Pliocene to Recent. 275 pp. (14 pis., n text-figs.). Clarendon Press, Oxford. 1941. The mollusca of a Norfolk Broad. /. Conch. 21 (8) : 224-243. 1951. Census of the distribution of British non-marine mollusca. Ibid. 23 (6 and 7) : 171-244 (8 pp. maps). ESMARK, B. 1883. Nyt Bidrag til Kundekaben om Norges Land og Ferskvands-Mollusker. Nyt. Mag. Naturv. 27 : 77-110. FERUSSAC, J. B. L. D'A. DE. 1819. Histoire naturelle generale et particuliere des mollusques terrestres et fluviatiles 2 : i-xvi + 1-96. Fol. Paris. 1823. Op. cit. (Nouvelle division des pulmones sans opercule) 2 : 96 a- z, a A. FISCHER, P. H. See Gassies & Fischer, 1856. GAIMARD, J. P. See Quoy & Gaimard, 1824. GASSIES, J. B. 1849. Table methodique et descriptif des mollusques terrestres et d'eau douce de I'Argenais. 209 [4] pp. (4 pis. col.). 8vo. Paris. GASSIES, J. B. & FISCHER, P. H. 1856. Monographic du genre Testacelle. Act. Soc. linn. Bordeaux 21 : 195-248 (2 pis.). GERHARDT, U. 1933. Zur Kopulation der Limaciden. Z. Morph. Okol. Tiere 27 (3) : 401-450 (n text-figs.). 1935. Weitere Untersuchungen zur Kopulation der Nacktschnecken. Ibid. 30 (2) : 297-332 (9 figs.). 1939. Neue biologische Untersuchungen an Limaciden. Ibid. 35 (2) : 183-202 (3 figs.). - 1940. Neue biologische Nacktschneckenstudien. Ibid. 36 (4) : 557-580 (8 figs.). GERMAIN, L. 1930. Mollusques terrestres et fluviatiles. Faune Fr. 21 : 447 pp. (13 pis. 470 figs.). GISTEL, J. VON N. F. X. 1848. Naturgeschichte des Thierreichs fur hohere Schulen. xvi + 216 [4] pp. (32 pis. col.). Fol. Stuttgart. (Mollusca pp. 165-174.) GRAY, J. E. 1821. A natural arrangement of mollusca, according to their internal structure. London Medical Repository 15 : 229-239. 1824. On the natural arrangement of the pulmonobranchous mollusca. Annals of Philo- sophy, Land. (N.S.) 24 : 107-109. 1840. A manual of the land and freshwater shells of the British Islands, with figures of each of the kinds. By William Turton (A new edition thoroughly revised and much enlarged by J. E. Gray) ix [i] + 324 pp. (12 pis. col., text illustrated). 8vo. London, i Apr. 1855. Catalogue of Pulmonata or air-breathing mollusca in the collection of the British Museum. Part i. [iv] + 192 pp. (text illustrated). 12 mo. London, 10-18 Apr. GREGG, W. O. 1944. Proterogyny in Deroceras panormitanum. Nautilus 58 : 6768. HAYWARD, J. F. 1954^. Agriolimax caruanae Pollonera as a Holocene fossil. /. Conch. 23 (12) : 403-404 (pi. 15). 19546. A further note on Agriolimax caruanae Pollonera. Ibid. 24 (i) : 21-22. HAZAY, J. 1881, Die Molluskenfauna von Budapest. Malakozool. Bl. (N.S.) 3 : 1-69 (pi. i, figs, i, 2). HELD, F. 1836. Aufzahlung der in Bayern lebenden Mollusken. Isis 29 (4) : 271-282. 1848-49. Die Landmollusken Bayerns. Jahresb. der K. Kreis-Landwirtschafts- und Gewerbs- Schule zu Munchen fur das Schuljahr, 1848-49 : 23. 4to. Miinchen. HESSE, P. 1926. Die Nacktschnecken der palaearktischen Region. Abh. Arch. Molluskenk. 2 (i) : 1-152 (2 pis.). Frankfurt. HEYNEMANN, D. F. 1861. Die nackten Schnecken des Frankfurter Gebiets, vornehmlich aus der Gattung Limax. Malakozool. Bl. 8 : 85-105. 1873. Ueber Geomalacus. Ibid. 21 : 25-36 (taf. i). See Koch & Heynemann, 1874. HOFFMANN, H. 1925. Zur Kenntnis der Testacellen. Arch. Molluskenk. 57 (5 and 6) : 171- 221 (taf. v, vi). HUTTON, F, W. 1879. Description of some new slugs. Trans, Proc, N.Z, Inst. 11 : 331-332, 220 BRITISH SLUGS JEFFREYS, J. G. 1862. British Conchology, Vol. I, land and freshwater shells, cxiv -f- 341 pp. (8 pis.). London. JENSEN, O. S. 1873. Indberetning om en i Sommeren 1870 foretagen Reise i Kristiania og Kristianssands Stift forat undersoge Land og Ferskvands-Molluskerne tilligemed Iglerne. Nyt. Mag. Naturv. 19 : 146-188 (pi. i). JOHNSTON, G. 1828. A few remarks on the class mollusca, in Dr. Fleming's works on British animals ; with descriptions of some new species. Edinburgh new Phil. Journ. 5 : 74-81. 1838. A list of the pulmoniferous mollusca of Berwickshire and North Durham. Berwick- shire Naturalists' Club 1 : 154156. KALENICZENKO, J. 1839. Necrologie : Notice sur le Professor J. A. de Krynicki. Bull. Soc. Imp. Nat. Moscou (i) : 25-33. 1851. Description d'un nouveau genre de limaces de la Russie meridionale. Ibid. 24 : 215-228 (pis. 5, 6). KENNARD, A. S. & WOODWARD, B. B. 1926. Synonymy of the British non-marine mollusca (Recent and Post-tertiary), xxiv -f 447 pp. London. KEW, W. 1902. On the mucous-threads of land-slugs. /. Conch. 10 (5) : 153-165. KOCH, C. & HEYNEMANN, D. F. 1874. Neue Nacktschnecken aus Turkestan. Jb. dtsch. malakol. Ges. 1 : 151-153. KUNKEL, K. 1916. Zur Biologie der Lungenschnecken Ergebnisse vieljdhriger Zuchtungen und Experimente. xvi + 440 pp. (48 text-figs., col. pi.). Heidelberg. LE COMTE, T. See Comte, T. le. LEHMANN, R. 1864. Neue Nacktschnecke aus Australien. Malakozool. Bl. 11 : 145-149. LESSONA, M. 1880. Molluschi viventi del Piemonte. Mem. Accad. Lincei (3) 7 : 317-380 (4 pis.). LESSONA, M. & POLLONERA, C. 1882. Monografia dei Limacidi Italiani. 73 pp. (3 pis.). 4to. Torino (preprint of Mem. R. Accad. Torino (2) 35 : 49-128, 1884). LEYDIG, F. 1876. Die Hautdacke und Schale der Gastropoden, nebst einer Uebersicht der einheimischen Limacinen. Arch. Naturgesch. 42 (i) : 209-292 (pis. ix-xvi). LINNAEUS, C. 1758. Sy sterna Naturae . . . Editio decima, reformata 1 [iv] -f 824 pp. 8vo. Holmiae. 1767. Sy sterna Naturae . . . Editio duodecima reformata 1 (2) : 533-1327. 8vo. Holmiae. LISTER, M. 1674. An extract of a letter of Mr. Martin Lister concerning the first part of his tables of snails, together with some queries relating to those Insecta, and the tables themselves. (Some general queries concerning land and freshwater Snails.) Philosophical Transactions, 9 (105) : 96-99 (pi. 2). 1685-97. Historiae sive synopsis methodicae conchyliorum, etc. 6 pts. Fol. Londini, 1685- 92 (-97)- LOCARD, A. 1895. Ipsa Draparnaudi conchylia. Etude sur la collection conchyliologique de Draparnaud au Musee Imperial et Royal d'Histoire Naturelle de Viennes. 190 pp. Paris. LOHMANDER, H. 1937- t)ber die nordischen Formen von Avion circumscriptus Johnston. Acta. Soc. Fauna Flora fenn. 60 : 90-112 (17 figs.). LUKIS, F. C. 1834. A notification of the occurrence, in the Island of Guernsey, of a species of Testacellus, and of some of its characteristics and habits, as observed there. Mag. nat. Hist. 7 : 224-226 (2 figs.). LUTHER, A. 1915. Zuchtversuche an Ackerschnecken (Agriolimax reticulatus Mull, und Agr. agrestis L.). Acta. Soc. Fauna Flora fenn. 40 (2) : 1-42 (2 text-figs.). MABILLE, M. J. 1867. Archives malacologiques ... I. Le genre Geomalacus. Rev. Mag. Zool. (2) 19 : 53-64. 1868. Des limaciens europeens. Rev. Mag. Zool. (2) 20 : 129-146. 1883. Sur quelques especes de mollusques terrestres. Bull. Soc. philom. Paris (7) 7 : 39-53- MAKINGS, P. 1959. Agriolimax caruanae Pollonera new to Ireland. /. Conch. 24 (10) : 354-356. MALM, A. W. 1868. Zoologiska observationer. V. Haftet. Skandinaviska Land-Sniglar, Lima- cina etc. (5 pis.). 8vo, Goteborg. (A preprint from Goteborgs VetenskSamh, Handl. 1870, 10 : 26-93.) BRITISH SLUGS 221 MILLET, P. A. 1843. Description de plusieurs especes nouvelles de mollusques de France. Mag. Zool. Paris. (2) (Classe V) Mollusques, 4 pp. (pis. 63, 64). MOQUIN-TANDON, A. 185556. Histoire naturelle des mollusques terrestres et fluviatiles de France, etc. 2 torn, (illust.). 8vo. Paris. MORCH, O. A. L. 1865. Quelques mots sur un arrangement des mollusques pulmones terrestres (Geophiles, Fer.) base sur le systeme naturel (suite). /. Conchyliol. 13 : 376-396. MORELET, P. M. A. 1845. Description des Mollusques terrestres et fluviatiles du Portugal. 8vo. Paris (July). MORTILLET, G. See Dumont & Mortillet, 1856. MtiLLER, O. F. 1774. Vermiumterrestrium et fluviatilium . . . Historia. 2. xxxvi -f 214 + [10]. 4to. Havniae & Lipsiae. NILSSON, S. 1822. Historia molluscorum sveciae terrestrium et fluviatilium breviter delineata. xx + 124 pp. 8vo. Lundae (1822, 1823). NORMAND, N. A. J. 1852. Description de six limaces nouvelles observees aux environs de Valenciennes. 8 pp. 8vo. Valenciennes, Nov. OKLAND, F. 1923. Arionidae of Norway. Skr. VidenskSelsk., Christ. 1922 (5) : 1-62 (pi. i, 46 figs., 5 maps). OLDHAM, C. 19420. Auto-fecundation and duration of life in Limax cinereo-niger. Proc. malac. Soc. Lond. 25 (i) : 9-10. 19426. Notes on Geomalacus maculosus. Ibid. 25 (i) : 10-11. ORBIGNY, A. C. V. M. D'. 1836-44. Mollusques. Histoire Naturelle des lies Canaries. (Webb & Berthelot) 2 (2) : 4-117 (pis. i-vii B). PHILIPPI, R. A. See Schultz, A. W. F. 1836-44. PHILLIPS, R. A. & WATSON, H. 1930. Milax gracilis (Leydig) in the British Isles. /. Conch. 19 (3) : 65-93 (pis. 1,2). PILSBRY, H. A. 1944. West American field slugs (Deroceras}. Nautilus 58 : 15-16. - 1948. Land Mollusca of North America (North of Mexico). Monogr. Acad. nat. Sci. Philad. (No. 3) : 2 i-xlvii + 521-1113 (figs. 282-585). PINT, N. 1876. Molluschi terrestri e d'acqua dolce viventi nel territorio d'Esino. Bull. Soc. Malac. Ital. 2 : 67-205 (2 pis.). PLATE, L. H. 1891. Studien uber opisthopneumone Lungenschnecken. Die Anatomic der Gattungen Daudebardia und Testacella. Zool. Jb. 4 : 505-630 (pis. 32-37). POLLONERA, C. 18870. Intorno ad alcuni Limacidi europei poco noti. Boll. Mus. Zool. Anat. comp. Torino 2 (21) 14 pp. (i pi). - 1887^. Specie nuove o mal conosciute di Avion europei. Atti Accad. Torino 22 (5) : 290- 313 (i pl-)- - 1889. Nuove contribuzioni allo studio degli Avion europei. Ibid. 24 : 623-640 (i pl.). 1891. Appunti di Malacologia. Intorno ai Limacidi di Malta. Boll. Mus. Zool. Anat. comp. Torino 6 (99) : 1-4 (2 figs.). 1882. See Lessona & Pollonera, 1882. QUICK, H. E. 1946. The mating process in Avion hortensis Ferussac and in Avion subfuscus Draparnaud. /. Conch. 22 (8) : 178-182 (8 figs.). - 1947. Avion ater (L.) and A. rufus (L.) in Britain and their specific differences. Ibid. 22 (10) : 249-261 (23 figs.). 1949. Synopses of the British fauna. No. 8. Slugs (Mollusca). (Testacellidae, Arionidae, Limacidae.}. 29 pp. (2 pis. 26 text-figs.). Linnean Society of London. 1950. The spermatophore of Milax sowerbii (Ferussac). /. Conch. 23 (4) : 111-112 (8 figs.). 1952. Rediscovery of Avion lusitanicus Mabille in Britain. Proc. malac. Soc. Lond. 29 (2 and 3) : 93- 101 ( J 5 fi gs.). - 1954. Journey to the Gugh< Highlands (Southern Ethiopia). 1948-49 : Gastropoda, slugs from the high mountains. /. Linn. Soc. (Zool.) 42 : 382-386 (17 figs.). QUOY, J. R. C. & GAIMARD, J. P. 1824. Voyage autour de monde . . . execute sur . . , I'Uranie et la Physicienne . . . Zoologie text. 671 pp. 4to. Paris, 222 BRITISH SLUGS RAFINESQUE-SCHMALTZ, C. S. 1815. Analyse de la nature, etc. 8vo. Palerme. 18200. Annals of Nature or Annual Synopsis of new genera and species of animals, plants etc. discovered in North America . . . First Annual Number, for 1820. i6pp. 8vo. Lex- ington. 18206. Monographic des coquilles bivalves fluviatiles de la Riviere Ohio, etc. Ann. gen. Sci. phys. 5 : 287-322 (3 pis.). RAZOUMOWSKY, G. DE. 1789. Histoire naturelle du Jorat et de ses environs, etc. 1. 8vo. Lausanne. (Mollusca pp. 266-278.) RECLUZ, C. A. 1845. Description des Mollusques terrestres et fluviatiles du Portugal, par A. Morelet. Rev. zool. 1845 : 306-312. Risso, A. 1826. Histoire naturelle des principales productions de I' Europe meridionale, et particulierement de celles des environs de Nice et des Alpes Maritimes. 4. 8vo. Paris. ROEBUCK, W. D. 1883. The British Slug List. /. Conch. 4 (2) : 38-43. 18840. A new variety of Avion ater. Ibid. 4 (5) : 146. i884&. Arion ater var. bicolor in West Gloucestershire. Ibid. 4 (7) : 217. 1885. New varieties of Limax arborum and Arion ater. Ibid. 4 (12) : 375-376. 1921. Census of the distribution of British land and freshwater mollusca. Ibid. 16 (6) : 165-212. ROISSY, A. F. P. M. DE. 1805. Histoire naturelle . . . des Mollusques, etc. (Suites a Hist. Nat. Buffon. redig. Sonnini.) 5 and 6. 8vo. Paris, An. 13. SANDBERGER, C. L. F. VON. 1880. Ein Beitrag zur Kenntnis der unterpleistocanen Schichten Englands. Palaeontographica (N.F. 7) 27 (2) : 83-104 (i pi.). SCHRANK, F. VON P. 1803. Fauna Boica. 8vo. Landshut. (Preface dated 8 Nov., 1802.) SCHRENK, . 1848. Uebersicht der Land und Susswasser Mollusken Livlands. Bull. Soc. imp. Nats. Moscou 21 (i) : 135-185. SCHULTZ, A. W. F. 1836-44. in Philippi, R. A. Enumeratio Molluscorum Siciliae, etc. 2 vols. 4to. Berolini, etc. SIEBERT, H. 1873. Die colorirten Tafeln des Lehmann' schen Werkes. NachrBl. dtsch. malakozool. Ges. (now Arch. Molluskenk.} 5 : 79-82. SIMROTH, H. 1885. Versuch einer Naturgeschichte der Deutschen Nacktschnecken und ihrer Europaischen Verwandten. Z. wiss. Zool. 42 (2) : 203-366 (pis. 7-11). 1887. Einige Bemerkungen, betreffend die Systematik der europaischen Nacktschnecken. NachrBl. dtsch. malakozool. Ges. (now Arch. Molluskenk.} 19 : 161-168. 1906. Ueber eine Reibe von Nacktschnecken die Herr Dr. Cecconi auf Cypern und in Palaestina gesammelt hat. Ibid. 38 (2) : 84-91. SORDELLI, F. 1870. Anatomia del Limax Doriae, Bourg., etc. (Descrizione succinta delle due species di Limax etc.). Atti Soc. ital. Sci. nat. 31 : 242-253 (i pi.). SOWERBY, G. B. 1820. The genera of recent and fossil shells . . . continued by G. B. Sowerby. pt. i. (first issue), pages not numbered. 8vo. London. STOKES, B. M. 1958. The worm-eating slugs Testacella scutulum Sowerby and T. haliotidea Draparnaud in captivity. Proc. malac. Soc. Land. 33 (i) : 11-20. See Barnes & Stokes, 1951. TAYLOR, J. W. 190207. Monograph of the land and freshwater mollusca of the British Isles. (Testacellidae, Limacidae, Arionidae). (pts. 8-13) : i-xx + 1-312 (25 pis. 275 figs, portraits). Leeds. THOMAS, D. C. 1944. Field sampling for slugs. Ann. appl. Biol. 31 (2) : 163-164. THOMPSON, W. 1840. Catalogue of the land and freshwater mollusca of Ireland. Ann. Mag. nat. Hist. 6 : 194-206. TURTON, W. 1831. A manual of the land and freshwater shells of the British Islands, viii + T-5 + [*6] PP- ( IO pl s -> col.). 8vo. London. See Gray, 1840. WATSON, H. 1915. Studies on the carnivorous slugs of South Africa, including a monograph of the genus Apera, and a discussion on the phylogeny of the Aperidae, Testacellidae, and other agnathous Pulmonata, Ann. Natal Mus. 3 (2) : 107-267 (pis. 7-24 ; 7 text figs.). BRITISH SLUGS 223 WATSON, H. 1930. Milax gracilis (Leydig) in the British Isles. Part 2. Anatomy. J . Conch. 19 (3) : 72-93 (pis. i and 2, text-figs. 8-23, map). 1943. Notes on a list of the British non-marine Mollusca. Ibid. 22 (3) : 53-72. See Phillips and Watson, 1930. WEBB & BERTHELOT. See Orbigny, 1836-44. WEIL, J. W. See Barnes & Weil, 1945. WEINLAND, D. F. 1876. Zur Wiechthierfauna der Schwabischen Alb. Jhr. Ver. vaterl. Naturk. Wurttemb. 32 : 234-358 (taf. 4). WESTERLUND, C. A. 1876. Neue Binnenmollusken aus Sibirien. NachrBl. dtsch. malakozool. Ges. (now Arch. Molluskenk.) 8 (8 and 9) : 97-104. WILLIAMS, J. W. 1888. The Shell-collector's handbook for the field, i-xii + 1-148 (14 figs.). 8vo. London. WILLIAMSON, M. H. See Cain & Williamson, 1958. WOLF, J. 1803-06. Die Wurmer. Sturm's Deutschlands Fauna, etc. (Abth. 6. Hft. i and 2). 8vo. WOLLASTON, T. V. 1878. Testacea Atlantica, etc. pp. xi + 588. 8vo. London. WOODWARD, B. B. See Kennard & Woodward, 1926. 224 BRITISH SLUGS INDEX Page agrestis, Agriolimax 163-4, J ^7' 168-70, 178-9 agrestis, Limax . . . 163, 168 Agriolimax 132, 144, 163-80, 181, 194, 198, 203 alba, Arion ater (var.) . . 139 alba, Arion intermedius (var.) . .125 alba, Arion subfuscus (var.) . . 133 alba, Milax sowerbyi (var.) . .158 albida, Arion ater (var.) . . 139 albina, Testacella haliotidea (var.) . 115 albina, Testacella maugei (var.) . . in albolateralis , Arion ater (var.) . .139 aldrovandi, Limax maximus (var.) . 192 allmani, Geomalacus maculosus (var.) . 120 Amalia ..... 149-50 Ambigolimax . . . . .194 andrewsi, Geomalacus maculosus (var.) 120 antiquorum, Limax . . . 187, 191 antiquorum, Limax flavus (var.) . . 184 aperta, Testacella maugei (var.) . . in arborum, Limax . . . .194 Arion . .120, 124-47, 153, 183, 203 Arionidae .... 119-47 Arioninae .... 119-47 ater, Arion 124, 126, 128, 130-31, 133-34, 136-7, 138-47 ater, Arion ater . . . 138-45 ater, Arion ater (var.) . . .139 ater, Limax .... 124, 138 aterrima, Arion ater (var.) . . 139, 145 aurea, Testacella maugei (var.) . . in aurea, Testacella scutulum (var.) . 117 aurigaster, Testacella . . .no babori, Amalia . . . 149-50 benoiti, Milax gagates (var.) . . 151 bettonii, Lehmannia marginata (var.) . 194 bevenoti, Agriolimax . . 172, 175 bicolor, Arion subfuscus (var.) . . 133 bicolor, Milax gagates (var.) . . 151 bilobatus, Limax . . . .170 Binneya . . . . . .119 bisulcata, Testacella . . . .118 breckworthiana, Limax flavus (var.) . 184 brunnea, Arion intermedius (var.) . 125 brunneus, Limax . . . .172 budapestensis, Amalia . . . 160 budapestensis, Milax 149, 155, 158-60, 160-63 caerulea, Arion hortensis (var.) . . 130 campestris, Limax . . . .172 canariensis, Limax . . . .164 Page Candida, Limax maximus (var.) . .192 Carinarion .... 124, 128 carinatus, Limax . . . .158 caruanae, Agriolimax 164,167, 169,173, 175-80 castanea, Arion ater (var.) . . 139 cellaria, Limax maximus (var.) . . 192 cerea, Limax tenellus (var.) . , 181 cereus, Limax . . . .180 cincta, Limax tenellus (var.) . . 181 cineracea, Agriolimax reticulatus (var.) 165 cinerea, Limax cinereoniger (var.) . 188 cinereus, Limax . . . .191 cinereofusca, Arion subfuscus (var.) . 133 cinereoniger, Limax 180, 183, 187-91, 192-3 circumscripta, Arion fasciatas (var.) . 128 circumscriptus, Arion . . .127 coccinea, Arion ater (var.) . . 139 coeruleus, Arion . . . .130 collingei, Amalia . . . .158 concolor, Limax maximus (var.) . .192 cyprius, Agriolimax . . . .164 Deroceras . . . . 163, 179 elongatus, Arion .... 130 empiricorum, Arion . . 124, 138, 145 Endodontacea .... 119-47 Endodontidae . . . . .107 europea, Testacella . . . .115 fasciata, Geomalacus maculosus (var.) . 120 fasciata, Limax maximus (var.) . .192 fasciatus, Arion 124-5, 127-9, 130-4, 144 fasciatus, Limax . . . .127 flavescens, Arion fasciatus (var.) . . 128 flavescens, Testacella haliotidea (var.) . 115 flavogriseus, Arion lusitanicus (var.) . 136 flavus, Arion . . . . .125 flavus, Limax 180, 184-7, 190, 196, 198, 200-1 fuliginea, Arion subfuscus (var.) . 133 fulva, Limax tenellus (var.) . . 181 fulvus, Limax . . . . .180 fuscatus, Arion . . . 133 fuscocarinata, Milax sowerbyi (var.) . 158 gagates, Limax . . . 149-50 gagates, Milax . . 149-55, 156-61 Geomalacus . . . 11924, 137 glauca, Lehmannia marginata (var.) . 194 globosus, Agriolimax . . . .175 gracilis, Limax . . . .160 grisea, Agriolimax laevis (var.) . .173 grisea, Arion fasciatus (var.) . . 138 grisea, Arion hortensis (var.) . .140 griseonigrescens, Testacella maugei (var.) 1 1 1 BRITISH SLUGS 225 Page griseorubescens, Testacella maugei (var.) in gughensis, Agriolimax . . .164 haliotidea, Testacella 108, 109-10, in, 114, 115-6, 117-8 hiberna, Avion ater (var.) . .139 hovtensis, Avion 124, 128-9, 130-2, 133-4, 137. M4 Hydrolimax . . . . .164 hyperboreus, Limax . . . .172 ichnusae, Amalia . . . .156 insularis, Amalia . . . .156 insularis, Milax . . 150, 155, 156-7 intermedia, Limax cinereus (var.) . 187 intermedius, Avion 124, 125-7, 129, 131, 134, 144 Kobeltia . . . . . .124 Krynickillus . . . . .164 lacustris, Agriolimax laevis (var.) . 173 laevis, Agriolimax 164, 167, 172-5, 178-9 laevis, Limax . . . . .172 Lehmannia . . . 193-200, 203 Limacidae .... 147-202 Limacinae .... 163-202 Limacus . . . . .180 Limax . 178, 180-93, 194, 196, 200-2, 203 lineatus, Avion . . . .187 Lochea ..... 124, 133 lusitanicus, Avion 124, 133, 135-8, 144-5 maculata, Agriolimax laevis (var.) . 173 maculata, Amalia . . . .158 maculosus, Geomalacus . 119, 120-4, 137 major, Testacella haliotidea (var.) . 115 marginata, Lehmannia 193, 194-7, 198-200, 202 marginatus, Limax . . . 158, 193-4 marginatus, Milax . . . .149 marginella, Avion ater (var.) . .139 maugei, Testacella . . 109-14, 115-18 maura, Limax cinereoniger (var.) . 188 maurus, Limax . . . .150 maximus, Limax 180, 183, 187-90, 191-3, 197 media, Avion atev (var.) . . .139 meditevranea, Amalia . . .150 megalodontes , Limax . . .184 melitensis, Lehmannia . . 193-4 Mesavion . . . . .124 Microarion . . . . .124 Milax . 149-63, 165, 178, 181, 183, 203 minimus, Avion . . . .125 minutus, Krynickillus . . .164 modiolifovmis , Limax . . . 202 nigva, Agriolimax veticulatus (var.) . 165 nigra, Avion hovtensis (var.) . .130 Page nigva, Testacella maugei (var.) . . in nigrescens, Avion lusitanicus (var.) . 136 nigrescens, Milax sowerbyi (var.) . 158 nigricans, Avion subfuscus (var.) . 133 nigvicans, Parmacella . . .150 nobrei, Avion . . . . 135 normalis, Avion intermedius (var.) . 125 nyctelius, Limax . . 180, 2001 Oleacinacea .... 108-18 Oleacinidae . . . . .107 olivaceus, Avion lusitanicus (var.) . 136 ovetea, Milax sowerbyi (var.) . 158, 160 ovnata, Limax cinereoniger (var.) . 188 pallidus, Limax . . . .170 panormitanus , Derocevas . . .179 panormitanus , Agriolimax . . . 164 Parmacella . . . . .149 Parmacellinae . . . 149-63 parryi, Amalia . . . 14950 pegoravii, Avion subfuscus (var.) . 133 pelhicida, Vitvina . . . .124 perkinsi, Agriolimax . . .175 plumbea, Avion atev (var.) . . .139 plumbea, Avion intermedius (var.) . 125 plumbea, Milax gagates (var.) . . 151 poirieri, Lehmannia . 193-4. 197-200, 201 poirieri, Limax .... 197 punctata, Limax cinereoniger (var.) . 188 vava, Milax gagates (var.) . . .151 veticulata, Agriolimax veticulatus (var.) 165 veticulata, Avion atev (var.) . .139 veticulatus, Agriolimax 163, 164-70, 171-2, 177-9 veticulatus, Limax . . . .164 vosea, Lehmannia marginata (var.) . 194 vufescens, Agriolimax veticulatus (var.) 165 vufescens, Avion hovtensis (var.) . 130 vufescens, Avion lusitanicus (var.) . 136 vufescens, Limax flavus (var.) . 184 vufofusca, Avion subfuscus (var.) . 133 vufus, Avion ater . 138, 139, 144, 145-7 vufus, Limax ..... 145 rupicola, Lehmannia marginata (var.) . 194 vusticus, Milax . . . .149 vutilans, Geomalacus maculosus (var.) . 120 scotti, Agriolimax . . . .164 scutulum, Testacella 108-10, in, 114, 116, 117-8 sevotinus, Limax . . . .180 similis, Amalia mediterranea (var.) . 150 sowerbyi, Limax . . . .158 sowerbyi, Milax . 149-50, 155, 158-60 subflavus, Avion .... 133 226 BRITISH SLUGS Page subfuscus, Arion 124, 129, 133-5, 136-7, 144. 147 subfuscus, Limax . . . 133 subrufa, Lehmannia marginata (var.) . 194 Succinea . . . . . .109 succinea, Arion ater (var.) . . .139 succinea, Arion subfuscus (var.) . 133 sulcatus, Arion . . . .145 sylvatica, Limax maximus (var.) . 192 tenellus, Agriolimax . . . .180 tenellus, Limax 180-4, 185-6, 188, 190 Testacella . . . 108-18, 203 Testacellidae .... 108-18 tetrazona, Limax maximus (var.) . 192 Page tigris, Limax maximus (var.) . . 192 tigrina, Lehmannia marginata (var.) . 194 trigona, Testacella haliotidea (var.) . 115 valentianus, Limax . . . 197-8 variegatus, Limax . . . .184 verkruzeni, Geomalacus maculosus (var.) 120 vinosa, Limax maximus (var.) . .192 violacea, Agriolimax reticulatus (var.) . 165 viridans, Testacella maugei (var.) . in virescens, Limax flavus (var.) . . 184 Vitrina . . . . . .178 Zonitacea .... 147-202 Zonitidae . . . . .107 Zonitoides . . . . .174 i. Geomalacus maculosus. 4. Arion hortensis. 6. Arion intermedius. 9. Arion subfuscus (young). 13 Testacella maugei. PLATE i The scale lines represent one centimetre. 3. Arion ater ater. 7. Arion intermedius (contracted) . 10. Testacella haliotidea. 12. Milax gagates (contracted). 2. Arion subfuscus. 5. Arion fasciatus. 8. Arion fasciatus (contracted). ii. A rion ater rufus (young). 14. Lehmannia poirieri. Bull. B.M. (N.H.) Zoo/. 6, 3 PLATE i PLATE 2 The scale lines represent one centimetre. 15. Agriolimax reticulatus. 16. Limax tenellus. 17. Agriolimax caruanae. 18. Lehmannia marginata. 19. Agriolimax laevis, 20. Milax budapestensis. 21. Lehmannia marginata var. tigrina. 22. Limax flavus. 23. Milax sowerbyi. (contracted) . 24. Limax maximus. 25. Limax cinereoniger. Bull. B.M. (N.H.) ZooL 6, 3 PLATE 2 16 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES (PISCES, CICHLIDAE) PART IV p. GREENWOOD BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 4 LONDON : 1960 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES (PISCES, CICHLIDAE) PART IV BY P. H. GREENWOOD Department of Zoology, British Museum (Natural History) Pp. 227-281 ; 21 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 4 LONDON: 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 4 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March, 1960 Price Twenty Shillings A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES (PISCES, CICHLIDAE) PART IV By P. H. GREENWOOD Department of Zoology, British Museum (Natural History) CONTENTS Page INTRODUCTION ........... 229 Haplochromis lacrimosus (Blgr.) ....... 230 Haplochromis pallidus (Blgr.) ........ 233 Haplochromis macrops (Blgr.) . . . . . . . .236 Haplochromis cinereus (Blgr.) . . . . . . . .239 Haplochromis niloticus nom. nov. ....... 243 Haplochromis martini (Blgr.) . .. . . . . .245 Haplochromis humilior (Blgr.) ........ 248 Haplochromis riponianus (Blgr.) ....... 252 Haplochromis saxicola sp. nov. ....... 256 Haplochromis theliodon sp. nov. ....... 260 Haplochromis empodisma sp. nov. . . . . . . .262 Haplochromis obtusidens Trewavas ....... 266 Haplochromis pharyngomylus Regan . . . . . .270 Haplochromis ishmaeli Blgr. . . . . . . . .275 SUMMARY ............ 280 ACKNOWLEDGMENTS .......... 280 REFERENCES ........... 280 INTRODUCTION IN previous parts of this series I have revised species-groups possessing common or related structural peculiarities and, in most cases, similar feeding habits. The present paper deals with a greater variety of structural and trophic types and many of the species show no obvious relationship to one another. With one exception, the species described below fall into three groups, namely, structurally generalized insectivorous species, specialized mollusc-eaters and species showing various degrees of structural and adaptational intermediacy between the other two groups. The exceptional fish, Haplochromis martini is a piscivorous predator ; it is included simply because of its overall resemblance to one of the insectivorous species described here. Formerly, some of the generalized species reviewed in this paper were synonymized with one of the mollusc-eaters (H. ishmaeli} ; the others are included because of various resemblances to species now resurrected and redefined. One such is Haplo- chromis cinereus, a species previously considered to be the extant representative of the ancestral type from which at least part of the present flock had evolved. These ZOOL. 6, 4 1 6 230 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES views on the central evolutionary position of H. cinereus are no longer tenable since the " species " thought to be H. cinereus was a complex of several distinct species, some more generalized than the others. Haplochromis cinereus, sensu stricto is, in fact, an anatomically somewhat specialized derivative from an even more generalized form. The most outstanding structural character in many of the species described below is an increase in the strength and size of the pharyngeal bones and musculature. As might be expected, these changes are reflected in the diet of the species, which usually include Mollusca as an important element in their food. Two species, H. ishmaeli and H . pharyngomylus , feed almost entirely on snails and bivalves. Those species with the pharyngeal mill in an intermediate stage of hypertrophy are able to deal with small molluscs and also with the tubicolous larvae of certain Trichoptera, an otherwise infrequent element in the food of insectivorous Haplo- chromis without strengthened pharyngeals. By crushing their molluscan prey within the pharynx these species stand in sharp contradistinction to the other groups of mollusc-eating Haplochromis in Lake Victoria. Species in this latter group remove the snail from its shell by holding the foot between the jaws and then levering the soft parts free before ingestion takes place (Greenwood, 19560 and 1957). Haplochromis lacrimosus (Blgr.) 1906 (Text-fig, i) Tilapia lacrimosa (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7) 17, 450 ; Idem, 1907, Fish. Nile, 515 ; Idem, 1915, Cat. Afr. Fish. 3, 234, fig. 154. Haplochromis cinereus (part) Regan, 1922, Proc. zool. Soc. London, 166. Lectotype. An adult male 76 mm. standard length (B.M. [N.H.] 1906.5.30.471) from Entebbe. Description, based on 36 specimens (including the lectotype and 10 paratypes) 66-0-97-0 mm. S.L. Depth of body 31-8-38-7 (M = 35-5) per cent of standard length, length of head 30-8-35-5 (M = 33-5) per cent. Dorsal head profile straight or slightly curved, sloping moderately steeply. Preorbital depth 13-6-18-0 (M = 15-5) per cent of head length, showing weak positive allometry with standard length ; least inter- orbital width 20-8-26-9 (M = 23-5), snout length 26-6-32-2 (M = 29-6) per cent. Diameter of eye 26-1-32-6 (M = 30-4) per cent of head, ratio of eye diameter to preorbital depth 1-5-2-3 (mode 2-0) ; depth of cheek 17-6-23-5 (M 20-8) per cent. Caudal peduncle 15-0-19-2 (M = 17-2) per cent of standard length, 1-2-1-8 (mode 1-5) times as long as deep. Mouth horizontal or almost so ; jaws equal anteriorly, the lower 31-4-41-3 (M = 37-1) per cent of head and 1-2-2-0 (mode 1-6) times as long as broad. Posterior tip of the maxilla extending to the vertical to the anterior orbital margin or slightly beyond, rarely not quite reaching the anterior orbital margin. Gill rakers variable, from moderately stout to slender ; 7-9 (mode 8), rarely 6 on the lower part of the first gill arch, the lowermost two or three reduced. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 231 Scales ctenoid ; lateral line with 31 (5), 32 (f.2i) or 33 (f.8) scales ; cheek with 2 or 3 (rarely 4) series ; 6 or 7 (rarely 5-|-) between the lateral line and the dorsal fin origin, 7 or 8 (rarely 6) between the pectoral and pelvic fin bases. Scales of the pectoral region small or moderate. Fins. Dorsal with 24 (f.J-9) or 25 (f.i6) rays, anal with n (f.ig), 12 (f.i6) or 13 (f.i), comprising XV-XVI, 8-10 and III, 8-10 spinous and branched rays for the fins respectively. Pectoral 82-5-100-0 (M = 88-5) per cent of head. Caudal truncate or subtruncate. Teeth. The outer row in both jaws is composed of unequally bicuspid, relatively slender and sometimes slightly recurved teeth. Rarely, a few unicuspid teeth may occur anteriorly in this row ; likewise a few posterolateral teeth in the upper jaw FIG. i. Haplochromis lacrimosus ; lectotype (from Boulenger, Fishes of the Nile). may be unicuspid. There are 40-60 (mode 54, modal range 48-54) teeth in the upper jaw. Teeth in the inner series are tricuspid and implanted at a slight, posteriorly directed angle (cf. H. cinereus where the inner teeth lie almost horizontally) ; there are 2 or 3 (rarely 4) rows of inner teeth in the upper jaw and 2 (less commonly 3, rarely 4) in the lower. Lower pharyngeal bone triangular and slender ; a few slightly enlarged but bicuspid teeth may occur in the median tooth-rows. Coloration. The colours of live fishes are unknown. Preserved material : Sexually active males. Ground colour yellowish-silver, chest dusky ; a dark lachrymal stripe is always present and in some specimens it may extend obliquely upwards through the eye and on to the nape ; there are usually two transverse bands across the snout. Two common patterns of body markings are known, (i) A large mid-lateral blotch 232 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES situated slightly posterior to the pelvic fin insertion and a mid-lateral stripe running from a point above the second anal ray on to the caudal fin ; a faint transverse bar is visible immediately posterior to the edge of the operculum and two others lie between the mid-lateral blotch and the origin of the posterior stripe. The blotch itself appears to be the intensified mid-portion of a vertical bar. (ii) Nine, close-set and ventrally ill-defined transverse bars on the flanks ; ventrally, the bars tend to run into one another so that the lower region of the flank is steely- grey. All fins, except the pelvics, hyaline, the upper part of the caudal sometimes weakly maculate, the mid-part dark ; anal with two or three ocelli ; pelvic fins black. Females and quiescent males. Ground colour greyish-silver, brownish above. Seven to nine faint transverse bars on the flanks, not reaching the ventral or dorsal outlines of the body. All fins hyaline. Distribution. At present, H. lacrimosus is known with certainty from Lake Victoria ; Pappenheim & Boulenger (1914) recorded a specimen from Lake Edward, but I have not been able to examine their material. Ecology : Habitat. No precise details are available for fishes already in the col- lections of the B.M. (Nat. Hist.) ; specimens collected by E.A.F.R.O. come from only two localities, both exposed, sandy beaches with the water depth less than 20 feet. Thus, it is impossible to generalize on the habitat preferences of H. lacri- mosus. Food. The stomach and intestinal contents of twenty fishes were examined ; with one exception (a fish from Entebbe) these specimens were caught at one time and at a single locality (Majita, Tanganyika Territory). The gut contents of the sixteen specimens containing food were varied. Twelve fishes contained fine sand- grains, bottom detritus (including fragments of plant epidermis and diatom frustules) and some Cladocera ; five contained remains of insect larvae (probably Diptera), one an adult dipteran, one the remains of a larval Povilla adusta Navas (Ephemerop- tera) and one an insect egg-mass. Two fishes yielded, besides insect fragments, the remains of an oligochaet worm, whilst two others each contained the foot and soft parts of a snail. From these scanty and topographically restricted data, H. lacrimosus should perhaps be classified as a bottom-feeding omnivore. Breeding. The breeding habits are unknown. Two of the smallest fishes (male and female, both 66-0 mm. S.L.) are adult. It seems possible that adult males reach a larger maximum size than do females. Affinities. Haplochromis lacrimosus is one of the structurally and ecologically generalized species of Lake Victoria. Its most striking and apparently diagnostic character is the markings of male fishes ; but, it must be stressed that coloration is known only from preserved material. In general appearance H. lacrimosus resembles H. pallidus (see p. 233) but the two species differ in several characters besides male coloration. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 233 Study material and distribution records Museum and Reg. No. Locality Collector Uganda B.M. (N.H.). 1906.5.30.471 Entebbe . Degen (Lectotype Tilapia lacrimosa) B.M. (N.H.). 1906.5.30.472-478 . . 1906.5.30.483-484 . Bunjako . ,, 1906.5.30.488-489 . Buganga . 1907.5.7.81-82 . Buddu Coast . Simon 1908.10.19.2-5 . Sesse Islands . Bayon 1959.4.28.24 . Entebbe Harbour . E.A.F.R.O. Tanganyika 1959.4.28.1-23 . Majita Haplochromis pallidus (Blgr.) 1911 (Text-figs. 2 and 3) Tilapia pallida (part) Boulenger, 1911, Ann. Mus. Genova (3), 5, 74; Idem, 1915, Cat. Afr. Fish. 3, 231-2. Labrochromis pallidus Regan, 1920, Ann. Mag. not. Hist. (9), 5, 45 (footnote). Haplochromis cinereus (part), Regan, 1922, Proc. zool. Soc. London, 166. This synonymy is tentative, as I have been unable to locate three specimens of T. pallida which Regan (1922) referred to Haplochromis guiarti (Pellegrin). Regan's genus Labrochromis, based on a skeleton wrongly identified as T. pallida, is discussed on page 275. Description, based on twenty specimens (including the holotype and four para- types) 43-74 mm. S.L. Depth of body 33'3~38-8 (M = 35-4) per cent of standard length, length of head 32-3-35-3 (M 34-1) per cent. Dorsal head profile straight or slightly curved, sloping at about 3O-4O. Preorbital depth 13-2-18-2 (M = 16-5) per cent of head ; least interorbital width 21-0-30-0 (M = 24-7) per cent. Snout as long as broad or slightly longer ; its length 29-1-33-4 (M = 31-2) per cent of head ; diameter of eye 26-1-33-4 (M = 29-5) per cent, ratio of eye diameter to preorbital depth 1-5- 2-3 (mode 1-7) ; depth of cheek 19-0-25-0 (M = 21-7) per cent of head. Caudal peduncle 14-2-18-5 (M = 16-2) per cent of standard length, 1-2-1-7 (mode 1-2) times as long as deep. Mouth horizontal ; jaws equal anteriorly, the lower 32-2-40-9 (M = 37-6) per cent of head and 1-4-1-8 (mode 1-6) times as long as broad. Posterior tip of the maxilla extending to the vertical through the anterior orbital margin or slightly beyond. Gill rakers short and stout (slender in one specimen), 7-9 (mode 9), rarely 10, on the lower part of the first arch, the lowermost 1-4 (or even 5) rakers greatly reduced. Scales ctenoid ; lateral line with 31 (f.6), 32 (1.13) or 33 (f.i) scales ; cheek with 2 or 3 (rarely 4) series ; 6 or 7 (occasionally 5) between the lateral line and the dorsal fin origin, 6 or 7 (occasionally 8) between the pectoral and pelvic fin bases. Scales on the chest rather small. 234 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Fins. Dorsal with 24 (f-5) or 25 (.15) rays, anal with n (f.3) or 12 (f.i7), com- prising XV-XVI, 8-10 and III, 8 or 9 spinous and branched rays for the fins respec- tively. Caudal truncate ; pectoral fin 78-0-87-0 (M = 81-0) per cent of head. Teeth. The outer row in both jaws is composed of unequally bicuspid, moderately stout teeth implanted erectly. In most fishes more than 67 mm. S.L. some unicuspid teeth occur postero-laterally in the upper jaw. The holotype (the largest specimen examined) has only weakly bicuspid teeth in the outer row. Three small specimens from near the Ripon Falls have somewhat more slender outer teeth than other specimens. There are 36-48 teeth in the upper outer series ; no clear-cut mode can be determined from the sample studied. The small and tricuspid inner teeth are implanted at a very slight angle and arranged in 3 (rarely 2) rows in the upper jaw and 2 or 3 rows in the lower. FIG. 2. Haplochromis pallidus ; holotype (from Boulenger, Ann. Mus. Genova). Lower pharyngeal bone triangular, usually slender but slightly thickened in three specimens. Most fishes have the two median rows of teeth slightly enlarged (especially the most posterior one or two pairs) ; in a few, all the pharyngeal teeth are slender. Coloration. The colours of live fishes are unknown. Preserved material. Adult males. Ground colour greyish, with faint traces of up to seven dark transverse bars on the flank and caudal peduncle ; branchiostegal membrane greyish. A distinct, vertical lachrymal stripe, continued at an angle, runs through the centre of the eye ; a very faint stripe runs from the posterior orbital margin to the angle of the preoperculum. Dorsal fin greyish, with dark lappets and a dark band along the basal two-thirds of the fin anteriorly, narrowing to the basal third or quarter posteriorly. Caudal dark proximally and along the mid-line. Anal hyaline, with two large, dead-white ocelli. Pelvics black, darkest laterally. Females and immature males. Ground colour greyish-silver, six or seven faint transverse bars on the flanks and, in some, two bars on the caudal peduncle ; a very faint lachrymal stripe. All fins hyaline. REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 235 Boulenger's (1911) account of preserved coloration differs somewhat from that given above, but as his material represented at least two and possibly three species, the discrepancies are not considered significant. Distribution. Known only from Lake Victoria, unless the locality " Jinja, Ripon Falls " for specimens nos. 1911.3.3.127-130 implies that these fishes were caught below the falls in the Victoiia Nile. Ecology. Habitat. The only precise bionomic data available are for those specimens collected by E.A.F.R.O. All these were from one locality, an exposed, shallow and sandy beach near Entebbe Airport. Food. Thirteen of the sixteen fishes examined had ingested matter in the stomach and intestine ; all these specimens were collected at one locality and at the same FIG. 3. Haplochromis pallidus; lower pharyngeal bone, oc- clusal view. 0-2 cm FIG. 4. Haplochromis ripon- ianus ; lower pharyngeal bone, occlusal view. FIG. 5. Haplochromis saxi- cola ; lower pharyngeal bone, occlusal view. time. Except for three fishes, all contained moderately large quantities of bottom debris (sand grains, diatom frustules and fragments of plant tissue) together with fragmentary insect larvae (? Diptera). The exceptional specimens contained only bottom detritus. Breeding. Haplochromis pallidus is a female mouth brooder ; the two smallest fishes examined (a male 54 mm. S.L. and a female 58 mm. S.L.) are both adult. Affinities. Haplochromis pallidus must be considered one of the many small and generalized species in Lake Victoria. Within this group it is extremely difficult to suggest phyletic relationships because the degree of inter-specific differentiation is so slight. In general appearance H. pallidus perhaps comes nearest to H. lacrimosus, from which species it is distinguished by a higher modal number of gill rakers (9 cf. 8), fewer and somewhat stouter teeth, a lower modal eye/preorbital ratio (1-7 cf. 2-0) and particularly, by differences in the preserved coloration of male 236 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES fishes. The nature of both oral and pharyngeal dentition suggests that a " pallidus " like species could have been ancestral to the adaptational grade at present represented by Haplochromis humilior (see p. 252). Study material and distribution records Museum and Reg. No. Locality Collector Uganda Genoa Museum (C.E. 12912) . Jmja . Bayon B.M. (N.H.). 1911.3.3.127-130 . Jinja, Ripon Falls . 1959.4.28.25-40 . Entebbe, Airport beach .. E.A.F.R.O. Haplochromis macrops (Blgr.) 1911 (Text-fig. 6) Haplochromis stanleyi (part) Boulenger, 1914, Cat. Afr. Fish. 3, 295. Haplochromis ishmaeli (part) Boulenger, 1914, torn. cit. 293. Tilapia macrops Boulenger, 1911, Ann. Mus. Geneva (3), 5, 73. pi. Ill, fig. i ; Idem, 1914, Cat. Afr. Fish. 3, 238. Haplochromis macrops (part, i.e. the species as described but excluding the tentative synonymy of Astatotilapia jeanneli Pellegrin), Regan, 1922, Proc. zool. Soc. London, 166. Description, based on forty specimens from Lake Victoria (including the holotype [Genoa Museum] and one of the paratypes) 66-91 mm. S.L. Depth of body 32-5-38-2 (M = 35-8) per cent of standard length, length of head 31-0-35-1 (M = 33-0) per cent. Dorsal head profile straight or slightly curved, sloping at a moderate angle (ca. 35-4O). Preorbital depth 11-5-16-3 (M = 14-2) per cent of head, least width of interorbital 26-6-32-2 (M = 29-7) per cent. Snout as broad as long or slightly broader, its length 26-6-31-0 (M = 29-0) per cent of head, diameter of eye 28-6-35-4 (M = 33'o) per cent, ratio of eye diameter to pre- orbital depth 2-0-2-9 (mode 2-3) ; depth of cheek 17-8-24-2 (M = 21-1) per cent of head. Caudal peduncle 14-1-18-4 (M = 16-8) per cent of standard length, 1-2-1-6 (mode 1-4) times as long as deep. Mouth slightly oblique, posterior tip of the maxilla extending to the vertical to the anterior orbital margin or even as far as the pupil. Lower jaw 38-0-42-5 (M = 39-5) per cent of head, 1-4-2-2 (modal range 1-7-1-8) times as long as broad. Gill rakers slender or, occasionally, relatively stout ; 8-n (mode 9) on the lower part of the first arch, the lower 1-4 rakers reduced. Scales ctenoid; lateral line with 30 (f.6), 31 (f.i7), 32 (f.i3) or 33 (f.3) scales; cheek with 2 or 3 series. Five or 6 (occasionally 7) scales between the lateral line and the dorsal fin origin ; 6 or 7 (occasionally 5, rarely 8) between the pectoral and pelvic fin bases. Scales on the pectoral region (relative to those on the ventral abdominal region) moderate to large. Fins. Dorsal with 24 (f.6), 25 (f-3o) or 26 (f.4) rays, anal with u (f-9), 12 (1.30) or 13 (f.i), comprising XV-XVII, 8-10 and III, 8-10 spinous and branched rays for the fins respectively. Pectoral 68-0-96-0 (M = 84-0) per cent of head. Caudal truncate. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 237 Teeth. The outer row of teeth in both jaws (except those situated postero-laterally in the upper) is composed mainly of bicuspid, moderately stout teeth. In certain fishes over 80 mm. S.L. some weakly cuspidate teeth may occur. H. macrops is, however, unusual in that the postero-lateral teeth in the upper jaw are generally tricuspid in fishes less than 85 mm. S.L. ; in larger individuals these teeth may be unicuspid. One aberrant individual has only tricuspid teeth in the upper, outer row ; the corresponding row in the lower jaw has a mixture of bi- and tricuspids, with the former predominating. There are 46-66 (mode 60, modal range 56-60) teeth in the upper, outer row. The inner tooth-rows are made up of tri- and unicuspid teeth arranged in 2 or 3 (rarely 4) rows in the upper jaw and I or 2 (less commonly 3) rows in the lower. Inner teeth are implanted so as to stand erect or with a very slight, posteriorly directed slope. FIG. 6. Haplochromis macrops ; holotype (from Boulenger, Ann. Mus. Genova). Lower pharyngeal bone triangular, not enlarged ; teeth fine and cuspidate. Coloration in life. Sexually active males. Ground colour dusky to intensely black. Dorsal fin black, lappets and margin of the soft part red, as are the spots and blotches between the branched rays. Anal dusky with a diffuse red flush becoming more intense distally : ocelli yellow. Caudal dusky, ventral half with a red flush. Pelvic fins dusky yellow. Quiescent males have a female-type coloration, but with yellow anal ocelli and red spots on the soft dorsal. Females. Ground colour greyish-yellow to silver-grey. Dorsal and anal fins light yellow, caudal yellowish. Colour of preserved material. Adult males. Dark blackish-brown, somewhat lighter, except in sexually active fishes, on the chest and branchiostegal membrane ; faint indications of six transverse bars on the flanks (generally not reaching the dorsal and ventral outlines) ; a dark but faint lachrymal stripe is visible in some specimens, 238 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES as is a faint dark bar along the preoperculum. Dorsal fin dark (especially along the basal third), lappets black. Caudal fin dark. Anal dark on the basal third to half, pale distally, with two or three ocelli. Pelvic fins black. Females. Brownish-yellow, some with eight or nine faint transverse bars on the flank and caudal peduncle. All fins hyaline but the caudal somewhat darker. Distribution. Definite records of Haplochromis macrops are available only from Lake Victoria although there is an indication that the species may also occur in the Lake Edward basin. I have examined material identified as H. macrops from Lake Edward (one specimen B.M. (N.H.) Reg. No. 1933.2.23.397 ; see Trewavas, 1933) and from rivers flowing into Lake Edward (see Poll, 1939 and Poll & Damas, 1939). Of these latter specimens (twelve in all) only one (R. G. Mus. Congo 31095, det. David, 1936), from Rutshuru, compares closely with the Victoria population of H. macrops. I hesitate to identify the remaining Congo Museum specimens from Rutshuru (Reg. Nos. 64888-64899), but the single specimen from the B.M. (N.H.) seems most closely allied to H. lividus Greenwood of Lake Victoria. I have not been able to study the two specimens from Lake Edward (now in Berlin) which Pappen- heim & Boulenger (1914) identified as H. macrops. The single specimen (R. G. Mus. Congo 31095) from the Edward basin now referred to H. macrops differs slightly from the generality of Victoria fishes in having a somewhat larger eye and longer lower jaw ; it is an adult female, 73-0 mm. S.L. The principal morphometric characters of this fish are : D* H* Po.% Io.% Eye% Snt. % Ck. % Lj. % C.R* Eye/Po. 32-8 31-4 15-2 30-4 37-0 30-4 21-7 43-4 15-0 2-4 * Per cent standard length. % Per cent head length. Dorsal XV, 9 ; anal III, 9 ; pectoral 91-3 per cent of head. In characters of dentition and squamation this fish is similar to those from Lake Victoria. Ecology. Habitat. The species is apparently confined to the shallow, sandy regions of the lake. Food. The predominating food organisms in the stomachs and intestines of twenty-four fishes (mainly from one locality, but caught on different occasions) are sub-imaginal Ephemeroptera ; two fishes had, however, fed almost exclusively on winged termites (Isoptera) and colonial blue-green algae. Typical bottom debris and sand-grains usually found in the guts of other insectivorous Haplochromis were not recorded. The occurrence of sub-imaginal or adult insects and planktonic blue- green algae, together with the absence of bottom debris, suggests that H. macrops feeds at the surface or in mid-water. Since the algae were not digested and did not constitute a major proportion of the ingested matter, they may be taken accidentally as the fishes dart after insect prey. Breeding. Haplochromis macrops is a female mouth-brooder. The smallest sexually active fish is a female 73 mm. S.L., the smallest adult male is 78 mm. S.L. It appears that males attain a larger maximum adult size than do females. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 239 Affinities. As with most of the structurally and trophically generalized Haplo- chromis of Lake Victoria, the detailed affinities of H. macrops are impossible to determine. The similarities existing between H . macrops and H. cinereus are discussed elsewhere (p. 242). The two species differ in several characters, particularly in their dentition and the larger eye/preorbital ratio of H. macrops (2-0-2-9, mode 2-3, cf. i '5-1-9, mode 1-8 for H. cinereus). Also, the gill rakers of H. macrops are finer and more numerous (mode 9) than in H. cinereus (mode 7). The large eye and shallow preorbital of H. macrops serve to distinguish the species from most other members of the " generalized species " group. Study material and distribution records Museum and Reg. No. Locality Collector Uganda Genoa Museum (C.E. 12928) . Bussu . Bayon (Holotype) B.M. (N.H.). 1911.3.3.137 . Bussu . Bayon (Paratype) B.M. (N.H.). 1911.3.3.114-115 . Jinja, Ripon Falls 1959.4.28.51-78 . Beach near Nasu Point . E.A.F.R.O. ,, 1959.4.28.79-84 . Buka Bay . ,, Tanganyika ,, 1959.4.28.85 . Mwanza, Capri Bay . ,, 1959.4.28.86 . Majita . ,, 1959.4.28.87 . Beach near Majita . ,, Haplochromis cinereus (Blgr.) 1906 (Text-fig. 7) Paratilapia cinerea Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 439 ; Idem, 1907, Fish. Nile, 478 ; Idem, 1915, Cat. Afr. Fish. 3, 344. Haplochromis stanleyi (part), Boulenger, 1915, torn, cit., 295. Tilapia lacrimosa (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 450 ; Idem, 1915, torn. cit., 234. Haplochromis cinereus (part), Regan, 1922, Proc. zool. Soc. London, 166. Haplochromis melanopus (part), Regan, 1922, op. cit. 165. Description, based on twelve specimens (including the holotype) 71-81 mm. S.L. Depth of body 34-6-39-0 per cent of standard length, length of head 30-8-37-3 (M = 34-7) per cent. Dorsal head profile straight or slightly curved, sloping at ca. 4o-5o. Preorbital depth 15-0-18-0 (M = 16-4) per cent of head length, least interorbital width 23-3-28-0 (M = 25-3) per cent. Snout as long as broad or slightly longer, its length 29-2-34-6 (M = 32-2) per cent of head, eye diameter 26-2-32-0 (M = 28-7) per cent, ratio of eye diameter to preorbital depth 1-5-1-9 (mode 1-8). Depth of cheek 20-8-26-0 (M = 23-0) per cent of head. Caudal peduncle 15-7-18-7 per cent of standard length, 1-4-1-6 times as long as deep. 240 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Mouth horizontal or very slightly oblique ; jaws equal anteriorly, the lower 34-6-41-3 (M = 37-7) per cent of head and 1-4-1-8 (mode 1-6) times as long as broad. Lips slightly thickened. Posterior tip of the maxilla extending almost to the vertical through the anterior orbital margin or as far as the eye. Gill rakers moderately stout, 7-9 (mode 7) on the lower part of the first gill-arch, the lowermost two or three rakers reduced. Scales ctenoid ; lateral line with 30 (f.i), 31 (f.i), 32 (f.4) or 33 (f.6) scales ; cheek with 3 or 4 series ; 5~6| scales between the lateral line and the dorsal fin origin ; 7 or 8 (rarely 6 or 9) between the pectoral and pelvic fin bases. Scales on the pectoral region small or, less frequently, moderate. Fins. Dorsal with 24 (f.8) or 25 (f.4) rays, anal with n (f.4) or 12 (f.8), comprising XV or XVI, 9 or 10 and III,' 8 or 9 spinous and branched rays for the fins respectively. Caudal truncate. Pectoral 72-5-92-0 (M = 80-2) per cent of head. FIG. 7. Haplochromis cinereus ; holotype (from Boulenger, Fishes of the Nile). Teeth. Teeth in the outer row of both jaws are weakly bicuspid or unicuspid, relatively stout and slightly recurved ; in most fishes there is an admixture of both types, with unicuspids predominating. There are 40-54 teeth in the upper, outer row. The inner series are composed of either unicuspid or bicuspid teeth ; less commonly there is a mixture of both types. A characteristic feature of the inner tooth-rows is the way in which the teeth are implanted obliquely so that the crowns point posteriorly. Three or 4 (rarely 2) inner rows occur in the upper jaw and 2 or 3 in the lower. The dentition of H. cinereus is unlike that of other species in the " generalized species ' ' group ; it is, indeed, typical of the dentition found throughout the large group of piscivorous predators. However, in other characters (syncranial architec- ture, feeding habits and body-form) H. cinereus is one of the generalized species (see p. 242). A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 241 Lower pharyngeal bone triangular, the dentigerous surface slightly broader than long. Teeth in the two median rows are slightly enlarged in five of the specimens examined, but are slender in the remaining seven fishes ; the other pharyngeal teeth are slender in all specimens. Coloration in life unknown. Colour of preserved fishes : Adult males. Ground colour greyish-brown ; chest and branchiostegal membrane dusky ; a faint lachrymal stripe. All fins hyaline, with very faint indications of dark maculae on the soft dorsal ; dorsal lappets dark. Anal fin with two or three whitish ocelli (orange surrounded by red in newly preserved material, according to Boulenger). Pelvics black except for a large light area extending over the distal half but not including the first branched ray. Females. Brownish, silvery grey ventrally, with eight or nine dark transverse bars on the flanks and caudal peduncle. All fins hyaline, the basal third to half of the caudal weakly maculate. Distribution. Known only from Lake Victoria. Ecology. No ecological data (except the locality) are available for the three specimens collected by Degen. The remaining nine fishes were caught in water less than 20 feet deep, over a sandy bottom on both protected and exposed shores. Food. The stomach and intestinal contents of ten specimens (from four localities) were examined. Of these, three were empty and the remainder yielded sand-grains, bottom detritus (including fragments of plant epidermis) and some larvae of dipterous insects. Breeding. The species is a female mouth-brooder. The smallest specimens available (a male and a female 71 mm. S.L.) are both sexually mature. As far as can be determined from this inadequate sample, both sexes reach the same adult size. Discussion of affinities and synonymy. In Regan's revision of the Lake Victoria Haplochromis (1922) the definition of H. cinereus was expanded to embrace a number of small and generalized or near-generalized forms previously recognized as distinct species. The first attempt to prune this complex was Lohberger's resurrection of H riponianus (see p. 252). Now, with more specimens available and some knowledge of Haplochromis in nature, it is clear that Regan's definition of H. cinereus must be narrowed considerably and that a further two species (H. lacrimosus and H. pallidus] should be resurrected. Amongst the group of anatomically and trophically un- specialized Haplochromis in Lake Victoria, H . cinereus is unusual because of its oral dentition (see p. 240). Relatively stout, clearly bicuspid teeth in the outer series of the jaws and erect tricuspid inner teeth are usual in the generalized species. The dentition of H. cinereus, on the other hand, shows a marked tendency for slender, unicuspid teeth to predominate in the outer rows ; the few bicuspid teeth present are weakly cuspidate. The inner teeth of H. cinereus are also atypical for the group in that the usual erect and tricuspid form is largely replaced by slender unicuspids implanted so as to point posteriorly. In fact, the dentition of H. cinereus is very like that of many predatory species. With so few specimens of H . cinereus known it is impossible to generalize on its feeding habits ; however, the gut contents of 242 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES seven fishes from four different localities do not even hint at the species being a piscivorous predator. At this point it should perhaps be stressed that my observations are confined to adult fishes 70-90 mm. long ; juveniles have still to be discovered. Because " Haplochromis cinereus " had become something of a dumping ground for any small Haplochromis species or specimen, the published information (Graham, 1929) on distribution and habitats can no longer be considered reliable. I have examined " H. cinereus " material collected by Graham and find that none of these specimens is referable to H. cinereus, sensu stricto. The bulk of this material is of undescribed species and will be dealt with in subsequent papers. Thus, Graham's remark that " the species (H. cinereus) is therefore widely distributed except in the deepest part of the lake " and Brooks' (1950, p. 159) elaboration of these data do not apply to H. cinereus but rather to the whole species-complex of generalized, bottom feeding Haplochromis in the lake. Haplochromis cinereus has been cited as representing the ancestral type from which the present species-flock could have evolved (Regan, 1922 ; Greenwood, 1951). For the reasons mentioned above this concept must be abandoned ; there are several other Haplochromis species still surviving in Lake Victoria which are structurally closer to the ancestral type, for example H. lacrimosus, H. nubilus or H. pallidus. The particular affinities of H. cinereus are difficult to determine. In gross anatomy and appearance H. cinereus does not differ markedly from the majority of small Haplochromis ; only when its dentition is considered does the difference appear striking. Like so many members of the generalized group, H. cinereus seems to be an independent offshoot from one of the basic stocks. In overall appearance and perhaps in at least some ecological requirements Haplochromis macrops is the one extant species most like H. cinereus. The two species are differentiated principally by the larger eye, shallower preorbital and the stouter, more numerous outer teeth of H. macrops. Haplochromis cinereus is not, as Regan suggested, closely related to H. ishmaeli. To stress this supposed relationship, Regan noted that six of the H. ishmaeli syntypes were actually specimens of H. cinereus. As a result of this present revision none of these specimens is still retained in H. cinereus. Haplochromis ishmaeli belongs to a distinct phyletic line, discussed more fully on pages 269 and 273. Study material and distribution records Museum and Reg. No. Locality Collector Uganda B.M. (N.H.). 1906.5.30.292 . Buganga . Degen (Holotype) B.M. (N.H.). 1906.5.30.350 . Entebbe . ,, 1906.5.30.482 . Bunjako ,, 1959.4.28.41 . Entebbe, Harbour . E.A.F.R.O. ,, 1959.4.28.48-50 . Entebbe, Airport beach . ,, Tanganyika ,, 1959.4.28.42-47 . Mwanza, Capri Bay . E.A.F.R.O. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 243 Haplochromis niloticus nom. nov. (Text-fig. 8) Tilapia bayoni Boulenger, 1911, Ann. Mus. Genova (3), 5; 72, pi. Ill, fig. 2 (nee Paratilapia bayoni [ = Haplochromis bayoni (Blgr.), see Regan, 1922, p. 176] Blgr., 1909, Ann. Mus. Genova (3), 4, 304, fig. Haplochromis humilior (part), Regan, 1922, Proc. zool. Soc. London, 169. Description, based on the holotype (100 mm. S.L.) and two other specimens 96 and 102 mm. S.L. FIG. 8. Haplochromis niloticus ; holotype (from Boulenger, Ann. Mus. Genova). The principal morphometric characters are given below : S.L. Depth* Head* Po. % lo. % Snt. % Eye % Ck. % T i / **J /o C.P.* 96-0 35'4 33-3 17-2 21-8 31-3 31-3 20-4 39'0 18-2 fioo-o 33' 33' 15-2 24-3 30-3 33'5 24-0 33'3 I7-5 IO2 -0 34-3 33-8 17-4 24-7 29-0 30-5 23-2 37'7 17-1 f Holotype. % Per cent, of head length. * Per cent, of standard length. Dorsal head profile sloping rather steeply (ca. 45-5o) and slightly curved. Mouth horizontal ; jaws equal anteriorly, the length /breadth ratio of the lower i-6-i-g. Posterior tip of the maxilla extending to the vertical through the anterior part of the eye. Lips not markedly thickened. ZOOL. 6, 4 17 244 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Gill rakers moderately stout ; 7 or 8 (the lowermost one or two reduced) on the lower part of the first arch. Scales ctenoid ; 32 in the lateral line ; cheek with 2 or 3 series. Six or 7 scales between the lateral line and the dorsal fin origin, 8 or 9 between the pectoral and pelvic fin bases ; scales on the pectoral region small. Fins. Dorsal with 25 (f.2) or 26 (f.i) rays, anal with 12, comprising XVI, 9-10 and III, 9 spinous and branched rays for the fins respectively. Pelvic fins with the first branched ray only slightly produced and extending to the origin of the anal. Caudal sub-truncate, scaled only on its proximal half. Teeth. The outer row in both jaws is composed of moderately slender, movably implanted and unequally bicuspid teeth ; the most posterior five or six teeth in the upper jaw are caniniform and stouter than those situated anteriorly and laterally. There are 65-70 teeth in the upper, outer series. Teeth in the inner series are small and tricuspid, and are arranged in four or five rows in the upper jaw and four in the lower. The interspace between the inner and outer series is very narrow. Lower pharyngeal bone triangular, slender or slightly enlarged, the two median rows of teeth relatively coarse in two specimens (including the holotype) and some- what more enlarged in the third. In the latter fish the next lateral row of teeth is also enlarged and the most posterior teeth of the median rows are sub-molariform. Osteology. A complete skeleton was prepared from one of the specimens caught at the same time as the three fishes described above. However, since H. niloticus is very similar to H. nuchisquamulatus (which also occurs in the Victoria Nile) it is difficult to confirm the specific identity of the skeleton. Apparently the sole diag- nostic osteological character is the lower pharyngeal bone, which is slender in H. nuchisquamulatus and slightly thickened in H. niloticus. The lower pharyngeal of the skeleton is that of H . niloticus and on this character alone the skeleton is referred to H. niloticus. In all other characters, except the oral dentition, the skeleton of H. niloticus resembles that of a generalized Haplochromis species. There are 14 + 16 vertebrae. Coloration of live fishes is unknown. The three preserved specimens are all ap- parently males (judging from the well-defined anal ocelli) and adult. Because most of the coloration is now lost (the fishes are a uniform brownish-grey) I quote the description given by Boulenger (1911) of the then newly preserved specimens. " Back dark olive to blackish, sides brassy yellow to coppery red ; a more or less distinct black bar below the eye ; dorsal and ventrals brown to black ; anal pink, blackish at the base, usually with two or three large orange ocellar spots ; caudal brown or blackish the lower third often pink." From this description the coloration of male fishes seems to be remarkably like that of H. humilior from Lake Victoria (see p. 250). Distribution. Haplochromis niloticus is known only from the Victoria Nile ; no information is available on its habitat or on feeding and breeding habits. Affinities. Haplochromis niloticus has been compared with two Haplochromis species from Lake Victoria. In his original description, Boulenger compared the A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 245 species with H. martini, whilst Regan (1922) considered H. niloticus to be conspecific with H. humilior. In my opinion H. martini and H. humilior are not closely related to one another and H. niloticus is not allied to either. The three species differ in several fundamental characters, especially in the nature of their dentition. Ad- mittedly, the somewhat enlarged median pharyngeal teeth of H. niloticus approach the condition found in some specimens of H. humilior, but the oral dentition of the two species is very dissimilar. The outer teeth are finer and more numerous in H. niloticus and there are more rows of inner teeth barely separated from the outer row. These same characters, together with a somewhat different arrangement of the jaw skeleton, serve to separate H. niloticus from H. martini. The nature of the dentition in H. niloticus suggests affinity with H. nigricans and H. nuchisquamulatus, particularly the latter. From H. nigricans, H. niloticus is distinguished by its more generalized neurocranium and dentary (see Greenwood, 19566) and its slightly enlarged lower pharyngeal teeth. From H. nuchisquamulatus, H. niloticus is again distinguished by having somewhat enlarged pharyngeal teeth and by a narrower interorbital region. The affinities of H. niloticus are not especially obvious ; the species is probably yet another slightly specialized side branch from the generalized Haplochromis stem. Study material and distribution records Museum and Reg. No. Locality . Collector Uganda Genoa Museum (C.E. 12932) . Kakindu, Victoria Nile . Bayon (Holotype T. bayoni Blgr. 1911) B.M. (N.H.) . 1911.3.3.124-5 . 1911.3.3.126 . (skeleton) Haplochromis martini (Blgr.) 1906 (Text-fig. 9) Tilapia martini, Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 449. T. martini (part), Idem, 1914, Cat. Afr. Fish. 3, 239, fig. 158. Haplochromis martini, Regan, 1922, Proc. zool. Soc. London, 171. Regan (1922) based his redescription of H. martini on three of the six syntypes, but did not indicate to what species he referred the remaining type specimens. However, amongst the material Regan identified as H. cinereus there are three type specimens of T. martini (B.M. (N.H.) Reg. No. 1906.5.30.466-468). Presum- ably it was Regan's intention to include these in the published synonymy of H. cinereus. The three specimens are not H. cinereus but are, in fact, Haplochromis martini. Lectotype. A brooding female 88 mm. standard length (B.M. (N.H.) Reg. No. 1906.5.30.465) from Bunjako, Uganda. Description, based on twenty-nine specimens (including the lectotype and four paratypes) 59-104 mm. S.L. 246 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Depth of body 30-8-38-0 (M = 34-4) per cent of standard length, length of head 31-0-38-1 (M = 35-4) per cent. Dorsal head profile very strongly decurved, the snout sloping at an angle of ca. $o-yo. Depth of preorbital 13-0-20-0 (M = 16-6) per cent of head, least interorbital width 20-4-26-8 (M = 24-1) per cent. Snout slightly longer than broad, its length 27-3-34-4 (M 30-2) per cent of head ; dia- meter of eye 29-4-37-5 (M = 31-7) per cent, depth of cheek 20-4-27-7 (M = 24-6) per cent. Caudal peduncle 15-3-20-6 (M = 17-3) per cent of standard length, 1-2-1-8 (mode 1-6) times as long as deep. Jaws equal anteriorly, the lower 38-4-45-8 (M = 42-6) per cent of head, 1-6-2-1 (modal range 1-7-2-0) times as long as broad ; mouth horizontal or slightly oblique. Posterior tip of the maxilla extending to the vertical through the pupil or, less commonly, only to the anterior part of the eye. Such a marked posterior extension of the maxilla is unusual in Lake Victoria Haplochromis and may be considered one of the diagnostic characters of H. martini. FIG. 9. Haplochromis martini ; lectotype (from Boulenger, Fishes of the Nile). Gill rakers on the first arch moderately stout (but rather slender in a few specimens), one or two of the uppermost often flattened and bifid ; 8 or 9, rarely 7 or 10 (mode 9) gill rakers on the lower part of the arch. Scales ctenoid; lateral line with 31 (.5), 32 (f-7), 33 (.9), 34 (.7) or 35 (f.i) scales ; cheek with 3 or 4 rows. Six to 8 scales between the lateral line and the dorsal fin origin, 7 or 8 (rarely 6) between the pectoral and pelvic fin bases. Fins. Dorsal with 24 (1.9), 25 (f.ig) or 26 (f.i) rays, anal with n (f.8), 12 (f.i-9) or 13 (f.2), comprising XIV-XVI, 8-10 and III, 8-10 spinous and branched rays for the fins respectively. Pectoral fin 73'5-865 (M = 82-0) per cent of head ; caudal truncate or sub-truncate. Teeth. In the outer row of both jaws the teeth are slender and mainly unequally bicuspid in fishes less than 85 mm. S.L. Above this size both weakly bicuspid and unicuspid teeth occur together. In fishes of all sizes a few tricuspid teeth are found postero-laterally in the upper jaw, an uncommon character in Lake Victoria Haplo- chromis (see H. macrops, p. 237). There are 46-76 teeth in the upper jaw (ill-defined modal range 68-70). A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 247 Tricuspid teeth predominate in the inner series although in one large fish (101 mm. S.L.) the inner teeth are all unicuspid. The inner series are arranged in 2 (less commonly 3) rows in the upper jaw and in i or 2 (less commonly 3) rows in the lower. Lower pharyngeal bone triangular, not enlarged and with slender, cuspidate teeth. Osteology. The neurocranium does not differ greatly from the generalized Haplo- chromis type, but the premaxilla has relatively longer dentigerous arms. Coloration in life. Sexually active males. Ground colour golden-yellow, shading to silvery- white, with three or four faint black blotches below the insertion of the dorsal fin, and a distinct coppery sheen on the nape and anterior part of the flank. Dorsal fin hyaline, with a pinkish flush. Distal half of the anal scarlet, the proximal half colourless ; anal ocelli orange-red. Caudal fin flushed with scarlet, especially intense on the distal half. Pelvics dark on the outer half, reddish-yellow mesially. Quiescent males golden-yellow shading to silvery-white ventrally ; a fairly distinct dark mid-lateral stripe and an interrupted upper stripe slightly below the insertion of the dorsal fin. Dorsal and caudal fins darkish, the anal and pelvics lighter ; anal ocelli yellow. Females and juvenile males. Ground colour and banding as above, the upper band usually broken into rather indistinct blotches. All fins light yellow, the caudal somewhat darker. Preserved material. Both sexes. Ground colour yellowish-silver to brownish, an intense, narrow mid-lateral black line extends from the upper angle of the oper- culum to the caudal peduncle and, in some specimens even on to the caudal fin ; a fainter, often interrupted black stripe runs mid-way between the upper lateral line and the dorsal fin base. In some fishes there are traces of a very faint interocular band. All fins hyaline ; in males the dorsal has dusky lappets and the pelvics are dark. Distribution. Haplochromis martini is known only from Lake Victoria. Specimens from Lake Edward once identified as Tilapia martini (Boulenger, 1914) were later referred to H. schubotzi (Regan, 1921). Ecology. Habitat. Available records (from eight localities) suggest that H. martini is restricted to water less than 40 feet deep, where the species is ubiquitous but nowhere common. The species has been found over both sand and mud bottoms, on exposed shores and in sheltered bays. There are some indications that it may not occur close inshore since the only record of H. martini in beach-operated seine nets came from an area (Majita) where the nets were shot about 300 yards off-shore. Food. Sixteen of the twenty-two specimens examined had ingested material in the stomach and intestines ; in three of these specimens, however, the contents were unidentifiable sludge. Eleven of the remaining thirteen fishes contained, as the exclusive or predominating food, the fragmentary remains of small fishes (identified in two cases as Haplochromis) ; one of these individuals had also fed on larval Diptera and another had eaten what appeared to be the foot and other soft parts of a snail. Another fish had fed only on larval Diptera and one was empty except for some small fish bones in the posterior intestine. 248 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Breeding. Haplochromis martini is a female mouth-brooder ; females reach sexual maturity at ca. 80 mm. S.L. ; no data are available for males. Both sexes appear to attain the same maximum adult size. Affinities. Because of its strongly decurved snout, large eye and the marked posterior extension of the premaxilla, H. martini is one of the more immediately recognizable species. Yet, despite these characters H. martini retains most of the fundamental features of a generalized species such as H. macrops. On the other hand, H. martini differs from members of this species group (and probably most other Lake Victoria species) in its bright yellow coloration. When attempting to assess the phyletic affinities of H. martini one is faced with these rather contradictory characters and with the fact that, despite its generalized dentition and body-form, H. martini can be a piscivorous predator. The majority of piscivorous Haplochromis in Lake Victoria are larger than H. martini and have elongate bodies and heads ; the teeth in these species are usually large and canini- form. Taking into account the various structural and ecological characters H. martini should perhaps be considered a superficially but trophically distinct branch from the generalized stem as represented, perhaps, by an H. macrops-like ancestor. Study material and distribution records Museum and Reg. No. Locality . Collector Uganda B.M. (N.H.). 1906.5.30.463-5 . Bunjako . Degen (Lectotype and paratypes) ,, 1906.5.30.466-8 . ,, ,, (Paratypes) ,, 1959.4.28.124-132 . Pilkington Bay . E.A.F.R.O. ,, 1959.4.28.138 . Old Bukakata Bay . ,, 1959.4.28.140 . Napoleon Gulf, near Jinja . ,, Tanganyika 1959.4.28.116-123 . Majita Kenya ,, 1959.4.28.133-137 . Off Port Southby . ,, 1959.4.28.139 . Beach below Usoma Lighthouse . Haplochromis humilior (Blgr.) 1911 (Text-figs. 10 and n) Tilapia humilior Boulenger, 1911, Ann. Mus. Genova (3), 5, 74, pi. Ill, fig. 3. Tilapia lacrimosa (part), Boulenger, 1915, Cat. Afr. Fish. 3, 234. Haplochromis desfontainesii (part), Boulenger, 1915, torn, cit., 303. Haplochromis nubilus (part), Regan, 1922, Proc. zool. Soc. London, 164. Haplochromis humilior (part), Regan, 1922, op. cit., 169. ? Paratilapia granti (part), Boulenger, 1915, torn, cit., 342. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 249 Lectotype. A male 90 mm. standard length from Kakindu, Victoria Nile, collected by Bayon (now in the collections of the Museo Civico di Storia Naturale, Genoa). Description, based on thirty specimens 65-90 mm. S.L. (including the lectotype and one paratype ; the second paratype [B.M. (N.H.) Reg. No. 1911.3.3.152] is very poorly preserved and although examined, is not included in the description). Depth of body 29-0-37-5 (M = 34-4) per cent of standard length, length of head 31-6-37-8 (M = 34-7) per cent. Dorsal head profile curved, sloping at an angle of 45-5o. Preorbital depth 13-6-17-9 (M = 16-3) per cent of head, least interorbital width 21-0-28-6 (M = 24-2) per cent. Snout as long as broad or slightly longer, its length 27-0-34-8 (M = 30-9) per cent of head, diameter of eye 27-0-32-5 (M = 30-3), depth of cheek 18-5-23-2 (M = 21-2) per cent. Caudal peduncle i5'2-i9-i (M = 17-4) per cent of standard length, 1-3-1-8 (mode 1-5) times as long as deep. FIG. 10. Haplochromis humilior, lectotype (from Boulenger, Ann. Mus. Genova). Mouth horizontal, the lower jaw often slightly shorter than the upper ; length of lower jaw 33-4-39-6 (M = 36-6) per cent of head and 1-3-2-2 (modal range 1-5-1-8) times as long as broad. Posterior tip of the maxilla extending to the vertical through the anterior orbital margin or somewhat beyond (to below the pupil in one specimen). Populations of H. humilior from different localities in the lake appear to have characteristic f acies which make any one population more or less readily identifiable ; unfortunately no means of quantifying these characters could be determined. Also, it has so far proved impossible, through lack of material, to decide whether Lake Victoria H. humilior differ from those inhabiting the Victoria Nile. Gill rakers short and stout, 6-8 (modes 6 and 7), rarely 9 on the lower part of the first arch. Scales ctenoid ; lateral line with 30 (.7), 31 (f.7), 32 (1.7) or 33 (.7) scales ; cheek with 2 or 3 (rarely 4) series. Six or 7 scales between the lateral line and the origin 250 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES of the dorsal fin, 6-8 (rarely 5 or 9) between the pectoral and pelvic fin bases. Scales on the pectoral region relatively small. Fins. Dorsal with 24 (f.6), 25 (f.2i) or 26 (.3) rays, anal with n (.4) or 12 (.24) comprising XV-XVI, 8-10 and III, 8 or 9 spinous and branched rays for the fins respectively. Pectoral 69'O-92'3 (M = 81-0) per cent of head. Caudal truncate. Teeth. The outer teeth in both jaws are moderately stout and unequally or, less frequently, sub-equally bicuspid. In fishes over 70 mm. S.L. some weakly bicuspid or even unicuspid teeth may occur ; there are 36-52 (modal range 46-48) teeth in the upper, outer series. Inner teeth are tricuspid and arranged in 2 or 3 (rarely 4) rows in the upper jaw and i or 2 (rarely 4) in the lower. A distinct space separates the inner and outer tooth series. Lower pharyngeal bone and teeth. The lower pharyngeal bone, although relatively stout is less massive than the bone in a specimen of H. ishmaeli or H. pharyngomylus 0-2 cm FIG. ii. Haplochromis humilior ; lower pharyngeal bone, occlusal view, from a Lake Victoria speci- men. 0-2 cm FIG. 12. Haplochromis theliodon ; lower pharyngeal bone, occlusal view. of the same size. Teeth in the two median rows are always enlarged, with the most posterior one or two pairs largest. There is some variation in the degree of enlarge- ment of the median teeth, which may be bicuspid, conical or even molariform ; no clear-cut correlation could be detected between, on the one hand, tooth-size and form and, on the other, the sex and size of the fish. There is, however, a tendency for larger individuals to have the coarsest median pharyngeal teeth. Sagittal sections through a number of bones (from fishes 65-70 mm. S.L.) suggest that both time and chance may influence the nature of the dentition, since unerupted replace- ment teeth are always molariform irrespective of the nature of the functional teeth which they underlie. In addition to the two median rows, the next lateral row on each side may also contain a number of molariform teeth. Coloration in life. Breeding males. Ground colour dark silvery-grey with intense dusky blotches on the head ; branchiostegal membrane dull black. A coppery flush extends over the cheek, operculum and flank as far as the origin of the anal fin. Anal and caudal fins light red, the colour becoming more intense along the A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 251 margins of both fins and the upper and lower posterior angles of the caudal ; two or three yellow anal ocelli. Dorsal fin dusky, with an orange-red margin to the soft part and red spots and bars between the rays of the posterior half of the spinous dorsal and over the entire soft part. Pelvics black. Quiescent males. General coloration as in females except that the pelvics are somewhat dusky and the un- paired fins have a pinkish flush ; anal ocelli are present. Females. Ground colour silvery-yellow. Dorsal, pelvic and anal fins pale yellow, the dorsal with red spots distributed as in males. Colour of preserved material. Males. Greyish, darker dorsally, the chest and branchiostegal membrane sooty ; in some specimens there are traces of seven or eight, narrow transverse bars on the flank and caudal peduncle. A lachrymal stripe, a pair of transverse stripes across the snout and, in some fishes, a broad band across the interorbital region are also visible, as are one or two bands on the nape. Dorsal fin variable, from hyaline to dusky. Anal hyaline in quiescent fishes and whitish in active individuals. Pelvics black (darkest in active fishes) but with a whitish overlay on the proximal half. Females silvery, some with very faint traces of vertical bars usually most obvious on the mid-lateral aspects of the flanks. All fins hyaline. Distribution. The species occurs in Lake Victoria and the Victoria Nile. Ecology. Habitat. In Lake Victoria the species is confined to shallow water over sandy beaches in both exposed and sheltered areas. No data are available for the riverine populations. Food. The gut contents of thirty-two fishes (from two localities) yielded identifi- able material. Of these specimens, twenty-four contained bottom debris (sand grains, plant fragments, diatom frustules and blue-green algae) together with remains of both larval and pupal insects (especially Diptera and Trichoptera, less frequently, Ephemeroptera) ; four fishes contained only insect remains and one only the diatom Melosira. Ten individuals had eaten, in addition to insects, both bivalves (unidentifiable) and gastropods (Melanoides sp. and Bellamy a sp.). These molluscan fragments were too finely divided to allow for any estimate of the number of animals eaten. Nevertheless it does seem, from this sample at least, that Mollusca are not a major element in the food of H. humilior. The large quantities of sand found with the remains of Trichoptera larvae in most fishes is of interest, particularly since the grains are small and of a remarkably uniform size. This suggests that the sand grains could be derived from the sand- grain cases made by certain caddis larvae. The moderately large pharyngeal mill of H. humilior may thus serve the dual purpose of crushing mollusc shells and the sand-grain cases of certain larval insects. Breeding. Nothing is known about the breeding habits of H. humilior. The smallest specimens available (a male and a female both 65 mm. S.L.) are adult ; the sexes apparently do not differ in maximum adult size attained. Affinities. Haplochromis humilior differs from the generality of small Haplochromis species in several characters, particularly in having a relatively massive lower 252 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES pharyngeal bone and in having a low modal number of gill rakers (6-7). Another pronounced difference lies in the tendency for the lower jaw to be shorter than the upper. In most other characters, H. humilior resembles H. pallidus, a species which shows incipient hypertrophy of the lower pharyngeal bone. Haplochromis humilior could be a more specialized off-shoot from an H. pallidus-like stem. Unlike many other presumed phyletic lines within the Victoria species flock, this one is not continued by one or more extant species. Haplochromis humilior does not appear to have any close relationship to the principal group of species with enlarged pharyn- geals, namely H. obtusidens, H. ishmaeli and H. pharyngomylus. Study material and distribution records Museum and Reg. No. Locality Collector Uganda Genoa Museum (C.E. 12910) . Kakindu, Victoria Nile . Bayon (Lectotype) B.M. (N.H.). 1911.3.3.152-3 . ,, (Paratypes) 1906.5.20.314 . Entebbe . Degen ,, 1959.4.28.88-107 . Beach near Nasu Point . E.A.F.R.O. ,, 1959.4.28.108 . Near Grant Bay . ,, ,, 1959.4.28.109-112 . Entebbe, Harbour Tanganyika 1959.4.28.113-115 . Majita Kenya 1909.11.15.38 . Kisumu Bay . Blayney-Percival Haplochromis riponianus (Blgr.) 1911 (Text-figs. 4 and 13) Pelmatochromis riponianus (part) Boulenger, 1911, Ann. Mus. Genova (3), 5, 69, pi. II, fig. 3 ; Idem, 1915, Cat. Afr. Fish. 3, 411, fig. 280. Haplochromis riponianus, Lohberger, 1929, Zoo/. Anz. 86, 222. Paratilapia serranus (part), Boulenger, 1915, torn, cit., 334. Paratilapia victoriana (part), Boulenger, 1915, torn, cit., 341. Haplochromis ishmaeli (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 446 ; Idem, 1915, torn. cit. 293. Haplochromis cinereus (part), Regan, 1922, Proc. zoo/. Soc. London, 166. On the basis of specimens in the Vienna Museum, Lohberger (1929) decided that Regan's views on the conspecificity of Pelmatochromis riponianus and H. cinereus could not be substantiated ; consequently he resurrected the former species as Haplochromis riponianus. I have not examined Lohberger's specimens but, from studying considerably more material than was available to either Regan or Lohberger, I can fully endorse the latter's action. Lectotype. A male 95-5 mm. standard length, from Jinja, Uganda, collected by Bayon (Genoa Museum, C.E. 12996). A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 253 Description, based on twenty-eight specimens (including the lectotype and two paratypes) 57-104 mm. S.L. One other paratype (B.M. (N.H.) Reg. No. 1911. 3 3 37) is n t included in the description. Depth of body 33-3-39-4 (M = 35-7) per cent of standard length, length of head 32-8-37-7 (M = 35-7) per cent. Dorsal head profile straight or very slightly curved, sloping at an angle of ca. 35-45. Preorbital depth 16-3-19-5 (M = 17-6) per cent of head, least interorbital width 23-0-28-1 (M = 25-1) per cent. Snout slightly longer than broad, its length 30-5-35-4 (M = 33-6) per cent of head ; diameter of eye 24-2-31-0 (M = 26-6), depth of cheek 19-2-25-0 (M = 22-7) per cent. Caudal peduncle 14-4-18-4 per cent of standard length, 1-2-1-7 times as long as deep (modal range 1-3-1-5). FIG. 13. Haplochromis viponianus ; lectotype (from Boulenger, Ann. Mus. Geneva) . Mouth horizontal or very slightly oblique, jaws equal anteriorly, the lower 33-4- 42-2 (M = 38-5) per cent of head and 1-3-2-0 (modal range 1-6-1-8) times as long as broad. Posterior tip of the maxilla extending to the vertical through the anterior orbital margin or almost so, occasionally to below the anterior quarter of the eye. Lips noticeably thickened but not produced into median lobes. Gill rakers on the first arch moderately stout in most fishes but slender in a few others, the lowermost one to three reduced and the pair nearest the epi-cerato- branchial angle often flattened and tri- or quadrifid ; 6-8 (mode 7) rakers on the lower part of the arch. Scales ctenoid ; lateral line with 30 (f.2), 31 (f.2), 32 (f.n), 33 (f-io) or 34 (f.2) scales, cheek with 3 or 4 series ; 5^-7 (rarely 5 or 8) scales between the lateral line and the dorsal fin origin ; 7 or 8 (rarely 6) between the pectoral and pelvic fin bases. Scales on the pectoral region moderate. 254 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Fins. Dorsal with 24 (.3), 25 (.24) or 26 (f.i) rays, anal with n (.3) or 12 (f.25), comprising XV-XVI, 9-10 and III, 8 or 9 spinous and branched rays for the fins respectively. Pectoral 69-0-88-5 (M = 78-5) per cent of head. Caudal sub-truncate. Teeth. The outer teeth in both jaws are relatively slender and slightly to strongly recurved ; the basic cusp pattern is unequally bicuspid, but the crowns are often so worn that the teeth appear to be weakly cuspidate or even unicuspid and bluntly incisiform. In some fishes over 80 mm. S.L. initially unicuspid teeth occur and may even be the predominating form in fishes more than 100 mm. S.L. The number of teeth in the upper, outer rows shows slight positive allometry with standard length ; there are 38-62 teeth in this row. Unicuspid and weakly tricuspid teeth are found in the inner rows ; often both types of teeth occur together, especially in fishes over 80 mm. S.L. The inner teeth are implanted at an angle and may be buried in the thickened oral mucosa (possibly a preservation artefact). In the upper jaw, the inner teeth are arranged in 3 or 4 (rarely 2) rows and in the lower in 2-4 (rarely i) rows. Boulenger's description (1911 and 1915) of the inner teeth as " minute " appears to stem from his being misled by the thickened oral epithelium which has hidden all but the tips of these teeth. Lower pharyngeal bone and teeth. The lower pharyngeal bone is triangular and in most specimens fairly stout. The relative degree to which the bone is enlarged is somewhat greater than that of the lower pharyngeal in H. humilior (see p. 250). In a few fishes, however, the bone is slender. This variation in stoutness is not entirely correlated with size since, although the smallest specimens have slender or but slightly thickened bones, some of the larger fishes do not have proportionately enlarged pharyngeals. The form of the teeth in the four median rows does show correlation with both the size of the individual and the stoutness of the bone. In fishes less than ca. 70 mm. S.L. some of the posterior teeth in the two median rows are enlarged ; in fishes 70- 80 mm. most of the median teeth are enlarged as are some or all of the teeth in the row on each side of the two median rows. Next, in specimens above 80 mm. S.L. (except those with slender pharyngeal bones) the teeth of these four rows are larger still, whilst those in the two median rows tend to be molariform. Finally, in fishes over 98 mm. S.L. some of the more lateral teeth are also molariform. Large indi- viduals with slender lower pharyngeals have only slightly enlarged median teeth, comparable with those of the smaller (< 90 mm.) specimens described above. Osteology. The shape of the neurocranium departs from the generalized type towards the form found in the elongated, piscivorous predators (as typified by H. mento). In H. riponianus, the slope from the anterior tip of the vomer to the base of the supraoccipital crest is less steep and more nearly straight than in the generalized type of skull ; also, the preorbital region in H. riponianus is relatively longer than in, for example, the skull of H. obliquidens or H. macrops. Coloration. The colours of live fishes are unknown. Preserved material. Sexually active males. Ground colour greyish-brown, chest dusky ; a dark lachrymal stripe and faint traces of four interrupted transverse bars on the flanks. Dorsal, caudal and anal fins hyaline, the anal with two or three A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 255 dead-white ocelli, usually arranged in a single row, but occasionally in two rows. Pelvic fins black. Immature males similar to females but with darker pel vies and small, distinct ocelli on the anal fin. Females silvery-grey shading to silver ventrally ; six to ten transverse bars of variable width (narrowest when most numerous) across the flanks and caudal peduncle ; sometimes, a very faint lachrymal stripe. All fins hyaline, the caudal occasionally maculate ; in a few specimens there are one or two, small, dead- white spots in the same position as the ocelli on the anal fin of males. Distribution. Lake Victoria and possibly the Victoria Nile. The latter locality is surmised from the data given in the original description of the species. In that paper (Boulenger, 1911), the locality is given only as Jinja, Ripon Falls, but in the introduction Boulenger implies that the entire collection, of which H. riponianus formed part, was from the Victoria Nile, that is below the Ripon Falls. Later (Cat. Afr. Fish. 3, 1914) the type locality is given more specifically as " Ripon Falls, Victoria Nile." Ecology. Habitat. The species is apparently confined to sand or rock substrates in the littoral regions of the lake ; it has been caught in both exposed and sheltered localities. Food. Stomach and intestinal contents of thirty-two fishes (from one locality) show that at least this population of H. riponianus was mainly insectivorous, although some fishes had also fed on Mollusca. Insect larvae (especially Trichoptera and Ephemeroptera) were found in every fish. As in H. humilior (see p. 251) large quantities of uniformly sized sand-grains were found in the intestines of most fishes. Since these grains closely resemble those forming the cases of certain Caddis-fly larvae found intact in some individuals, it is suggested that the sand was derived, at least partly, from crushed cases. In addition to the main contents listed above, eight fishes contained relatively large quantities of crushed bivalve shells (? Corbicula sp.) ; a few fragments of gastropod shells were also found in these individuals. Breeding. No information is available on the breeding habits of H. riponianus. The smallest sexually mature fish is a female 84 mm. S.L. ; the smallest adult male is 86 mm. S.L. Affinities. In combination, the oral and pharyngeal dentition set H. riponianus apart from other Lake Victoria species. In appearance and in possessing a similar oral dentition, H. saxicola (see p. 257) appears to be the nearest relative of H. riponianus although it differs in having a relatively slender pharyngeal bone with only the two median tooth rows slightly enlarged. The two species, which differ somewhat in their ecological relationships, may represent separate adaptive lines derived from a common ancestor. Haplochromis pallidus seems structurally suited for consideration as the extant representative of the presumed ancestor. (See also p. 235.) Although the lower pharyngeal bone and its dentition are similar in H . humilior and H. riponianus and although both species have very similar ecological require- 256 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES ments, the species show quite marked divergence in body-form (especially head shape) and oral dentition. What relationships there are probably lie far back in their phylogenetic history ; certainly the two species cannot be considered members of a recently evolved phyletic line. Boulenger's view (1911) that H. riponianus and H. microdon are closely related can no longer be held ; the two species have very different phylogenies, as witnessed by several anatomical and osteological characters, and equally distinctive habits (Greenwood, 1959). Boulenger was undoubtedly misled by the supposedly small teeth of both species (Greenwood, op. cit., and above) whereas in fact the teeth of H. riponianus are not minute and are of a markedly different form from those of H. microdon. Study material and distribution records Museum and Reg. No. Locality - Collector Uganda Genoa Museum (C.E. 12996 . Jinja, Ripon Falls . Bayon (Lectotype) B.M. (N.H.). 1911-3- 3- 37-39 1911.3.3.24 1906.5.30.280 . Entebbe . Degen 1906.5.30.394 . Bunjako . (Paratype H. ishmaeli) 1929.8.13.1 . Entebbe . Hoare 1959.4.28.141-157 . Entebbe, Airport beach . E.A.F.R.O. 1959.4.28.158 . Hannington Bay . ,, ,, 1959.4.28.159 . Buka Bay . ,, 1959.4.28.160-162 . Entebbe, Harbour . Lake Victoria, Locality Unknown 1901.6.24.86-87 . . Sir H. Johnson 1928.5.24.136-138 . . M.Graham Haplochromis saxicola,, sp. nov. (Text-figs. 5 and 14) Pelmatochromis riponianus (part) Boulenger, 1911, Ann. Mus. Genova (3), 5, 69 ; Idem, 1915, Cat. Afr. Fish. 3, 411. Haplochromis cinereus (part) Regan, 1922, Proc. zool. Soc. London, 166. Holotype. A female in-o + 25-0 mm. long (B.M. (N.H.) Reg. No. 1959.4.28.249) from Ramafuta island, Uganda. Description, based on twenty-seven specimens (including the holotype of the species and two paratypes of Pelmatochromis riponianus) 106-123 mm. S.L. Depth of body 34-8-42-5 (M = 37-8) per cent of standard length, length of head 35-3-42-5 (M = 37-8) per cent. Dorsal head profile straight or gently curved, sloping at ca. 3O-4O. Preorbital depth 15-2-19-0 (M = 17-6) per cent of head, A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 257 least interorbital width 23-9-29-8 (M = 26-9) per cent. Snout slightly longer than broad or, less commonly, as long as broad, its length 34-0-41-8 (M = 37-5) per cent of head ; diameter of eye 20-4-26-8 (M = 24-4), depth of cheek 21-4-26-2 (M = 23-9) per cent. Caudal peduncle 13-1-17-7 (M = 15-4) per cent of standard length, i-o- 1-5 (mode 1-3) times as long as deep. Mouth horizontal or very slightly oblique ; lips variably thickened but always noticeably enlarged. Lower jaw 39-7-46-5 (M = 43-0) per cent of head and 1-5- 1-9 (mode 1-8) times as long as broad. Posterior tip of the maxilla reaching or almost reaching the vertical through the anterior orbital margin, sometimes extending to below the eye. Gill rakers moderately stout, the uppermost two or three either slender or divided FIG. 14. Haplochromis saxicola ; holotype. Drawn by John Norris Wood. and somewhat flattened, the lowermost one to three usually reduced ; 7-9 (mode 8) rarely 6 on the lower part of the first arch. Scales ctenoid; 31 (f.2), 32 (1.15), 33 (f-9) or 34 (f.i) in the lateral line ; cheek with 3 or 4 series. Six or 7 (rarely 5) between the lateral line and the dorsal fin origin ; 6-8 (rarely 5 or 9) between the pectoral and pelvic fin bases. Scales on the pectoral region small. Fins. Dorsal with 24 (1.4), 25 (f.2o) or 26 (f.2) rays, anal with n (f.i), 12 (f.24) or 13 (f.i), comprising XIV-XVI, 9 or 10 and III, 8-10 spinous and branched rays for the fins respectively. One specimen has XVI, 6 rays in the dorsal and another II, 8 in the anal. Pectoral 67-8-81-0 (M = 74-5) per cent of head. Caudal truncate or sub-truncate. Teeth, In the outer series of each jaw the teeth are slender, recurved and generally unicuspid, but a few weakly bicuspid teeth may occur in fishes less than 115 mm. S.L. 258 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES The crowns of the teeth are often worn so as to assume a bluntly incisiform shape. There are 52-68 teeth in the outer row of the upper jaw (modal range 60-62). The inner tooth rows are composed of obliquely implanted and either unicuspid or both uni- and tricuspid teeth arranged in 2 or 3 (rarely 4 or 5) rows in each jaw. The dentition of H. saxicola bears a very strong resemblance to that of H. ripon- ianus, the major difference lying in the higher percentage of primarily unicuspid teeth in H, saxicola. However, it must be borne in mind that no available specimens of H. riponianus are as large as even the smallest H. saxicola. Lower pharyngeal bone and teeth. In ninety per cent of specimens examined, the lower pharyngeal bone is not enlarged ; in the remaining ten per cent it is somewhat thickened and resembles the lower pharyngeal of H. riponianus. The two median tooth rows are slightly enlarged in all specimens and may even be molariform in those individuals with enlarged lower pharyngeal bones. Very exceptionally, the median teeth are not noticeably larger than the more lateral ones. Osteology. The neurocranium of H. saxicola is very like that of H. riponianus (see p. 254) ; the premaxilla is, however, distinctive for its noticeably arched dentigerous arms which impart a characteristic peak to the antero-medial part of the bone. Coloration in life. Sexually active males. Ground colour dark grey-green, some scales on the flank with golden centres ; chest and branchiostegal membrane black ; a coppery-red flush on the operculum and flanks. Dorsal fin dark, lappets and maculae red ; caudal blue-grey with red posterior and ventral margins ; anal fin blue-grey, suffused with pink, especially the distal margins, ocelli bright yellow ; pelvic fins black. Quiescent males pale silvery-blue, almost grey dorsally ; some flank scales with golden centres, pectoral region silver. Dorsal, caudal and pelvic fins hyaline, the dorsal with pale red lappets and margin to the soft part, pelvics dusky on the lateral half, hyaline mesially. Females. Ground colour golden-grey ; all fins hyaline. Preserved material. Adult males (probably sexually active). Ground coloration dark grey-black, especially dark on the head, dorsal aspects of the flanks and along the ventral body surface. Lips light grey ; branchiostegal membrane black. Dorsal fin dark, with light maculae on the soft part. Anal fin black on the basal third, the distal two-thirds yellowish, with two grey ocelli. Caudal greyish ventrally, yellow around the margin. Females, juvenile and some apparently adult (? quiescent] males. Brownish-yellow ground coloration, darkest on the dorsal surface of the head ; four to seven, fairly narrow dark transverse bars on the flanks and caudal peduncle. Dorsal and caudal fins greyish, all other fins colourless, except for the black pelvics in adult males. Distribution. The species is known definitely from Lake Victoria but may also occur in the Victoria Nile. This locality is suggested by the presumed provenance of two P. riponianus paratypes now considered to be H. saxicola. For further dis- cussion, see under " Distribution " in H. riponianus, p. 255. Ecology. Habitat. The habitats in which H. saxicola has been caught are rather more varied than is common for most Lake Victoria Haplochromis. The species A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 259 occurs in shallow water over exposed sand or shingle beaches, amongst dense plant stands near rocky shores and over shingle and small boulders in water 10-30 feet deep. Finally, one specimen was caught more than a mile off-shore in nets set on a mud bottom in water ca. 180 feet deep. The most consistent factors in all these habitats (except the latter) are the hard substrates of sand, rock or shingle. Food. Twenty-six fishes (from nine different localities) were examined ; of these, only twelve contained any ingested material in the stomach or intestine. Sand grains and small pebbles (ca. 2 mm. in diameter) were found in eleven individuals ; fragmentary insect larvae (Diptera, probably chironomids) formed the principal food in eight fishes and snails in one other. All specimens contained a few frag- mentary Ostracoda and one, an almost entire larva of the boring May-fly Povilla adust a. These very unsatisfactory data suggest that H. saxicola is a bottom feeder which preys on various invertebrates, particularly insect larvae. Breeding. No information was obtained on the breeding habits of H. saxicola ; the smallest specimen (106 mm. S.L.) is sexually mature. Both sexes appear to reach the same maximum adult size. Affinities. In general appearance, oral dentition, neurocranial morphology and in certain ecological characters, Haplochromis saxicola closely resembles H. riponianus. The most striking difference between the species is the typically slender lower pharyngeal bone of H. saxicola and the stouter pharyngeal bone and teeth of H. riponianus. The mean snout and lower jaw proportional lengths for H. saxicola are somewhat greater than those for H. riponianus. If, however, these characters are taken for both species and plotted against standard length, no marked discontinuity is observed ; in all morphometric characters H . saxicola could well be large H . riponianus. The differences in the lower pharyngeal bones and dentition, however, are against this interpretation, the more slender bone with but slightly enlarged median teeth characterizing the larger specimens described as H. saxicola. In no Haplochromis species with enlarged pharyngeals does the bone and its dentition become less coarse with growth. Indeed, the reverse is usual (see Greenwood, 1959 and p. 277). Study material and distribution records Museum and Reg. No. Locality Collector Uganda B.M. (N.H.). 1959.4.28.249 . Ramafutalsl. . E.A.F.R.O. (Holotype) 1911.3.3.39 . Jinja, Ripon Falls . Bayon (Paratype of Pelmatochromis riponianus) Genoa Museum (A paratype of P. riponianus) . ,, ,, Bayon B.M. (N.H.). 1959.4.28.250-255 . Ramafuta Island . E.A.F.R.O. 1959- 4 -28 .256-267 . 1959.4.28.268-270 . Beach near Nasu Point . 1959.4.28.271-272 . Jinja . 1959-4-28.273 o 4' S., 33 14' E. ZOOL. 6, 4 18 Depth* Head* Po. % lo. % Snt. % Eye % Ck. % L.j. % C.P.* 37*3 37'3 17-9 25-0 32-2 25-0 21-4 39-3 15-3 35'7 36-9 16-2 25-8 32-2 26-8 25-8 38-7 14-3 36-6 36-6 16-1 22-6 35-5 24-2 22-6 38-7 15-4 38-4 33'7 I7-3 2 4 -I 33'5 24-1 2 4 -I 38-0 I5-I 38-6 36-3 18-1 25-0 34'4 25-0 24-4 37'4 14-8 38-2 36-5 18-5 24-6 33-8 24-6 27-0 38-4 15-2 36-8 33'7 17-2 25-0 34'4 26-6 25-0 39'0 13-7 260 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Haplochromis theliodon sp. nov. (Text-figs. 12 and 15) Holotype. A fish 95-0 + 20-0 mm. long (B.M. (N.H.) Reg. No. 1959.4.28.163) from Majita, Tanganyika Territory. Description, based on seven specimens (including the holotype) 79-95 mm. S.L. The principle morphometric characters are summarized below : S.L. 84-0 85-0 86-0 88-0 89-0 * Per cent, of standard length. % Per cent, of head length. Dorsal head profile straight except for a slight concavity above the eye, sloping fairly steeply ; snout longer than broad. Mouth slightly oblique ; lips, especially the upper, thickened. Lower jaw 1-5-1 -8 times as long as broad. Posterior tip of the maxilla almost reaching the vertical to the anterior orbital margin. Gill rakers variable, from stout (the commonest) to relatively slender ; 7 (f.i), 8 (f.5) or 9 (f.i) on the lower part of the first arch. The lowermost two to four rakers may be reduced. Scales ctenoid ; lateral line with 31 (f.2), 32 (3) or 33 (f.2) scales ; cheek with 3 or 4 series. Six to 9 scales between the lateral line and the dorsal fin origin, 7-9 between the bases of the pectoral and pelvic fins. Scales on the pectoral region small and deeply embedded. Fins. Dorsal with 25 rays, anal with 12, comprising XV-XVI, 9 or 10 and III, 9 spinous and branched rays for the fins respectively. Pectoral 64-3-72-0 (M = 69-5) per cent of head. Caudal truncate or sub-truncate. Teeth. In six of the seven specimens examined, the outer row in both jaws is composed of relatively stout, unequally bicuspid teeth, except for the most posterior pair in the upper jaw which are unicuspid. In the seventh specimen, the anterior and postero-lateral teeth are unicuspid and the lateral teeth unequally bicuspid. There are 36-46 outer teeth in the upper jaw. The inner teeth are small and tricuspid (with a few unicuspids in the exceptional specimen mentioned above) and are arranged in 3 or 4 rows in the upper jaw and 3 or 4 (rarely 2) in the lower. Inner teeth are implanted vertically or somewhat obliquely. Lower pharyngeal bone and teeth. The lower pharyngeal bone is thickened and moderately stout ; that is to say, it is comparable with the lower pharyngeal in most H. humilior but finer than that in H. obtusidens (see p. 267). A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 261 Five of the seven specimens examined have the four median tooth rows composed of large and molariform teeth, whilst the next lateral row of each side contains enlarged but clearly cuspidate teeth. In the exceptional fishes, only the median pair of tooth rows contains molariform elements and only one lateral row on each side has enlarged teeth. Coloration. The colours of live fishes are unknown. Preserved material. Quiescent male. Ground colour dusky bronze, with a very faint mid-lateral stripe, six interrupted transverse bars and a distinct lachrymal stripe. Dorsal fin dark, lappets pale ; caudal dark except for its pale distal margin ; anal pale (? orange) with four, small, white ocelli arranged in a single row. Pelvics black on the outer two-thirds, lighter medially. Females and immature males uniformly light brown except for the darker dorsal head surface and nape ; a dark FIG. 15. Haplochromis theliodon ; holotype. Drawn by Miss G. Osterritter. mid-lateral stripe of variable intensity is continuous in some specimens but interrupted in others above and slightly anterior to the anal fin ; when inter- rupted, the line is thickened to form a black blotch above the pectoral fin. A fainter, continuous dorsal stripe runs above the upper lateral line ; there are traces of seven to nine, variously interrupted transverse bars on the flanks but not on the caudal peduncle. Dorsal and anal fins hyaline, darker between the branched rays. Caudal maculate on the upper half, immaculate below. Pelvic fins hyaline in females but with the outer half dusky in males. Distribution. Known only from Lake Victoria. Ecology. Habitat. The few specimens known came from two localities, both exposed, shallow and sandy beaches. Food. Analysis of stomach and intestinal contents from all seven specimens indicate that H. theliodon is a bottom feeder with a varied diet including small fishes (cichlids, 20-25 mm. long), Gastropoda (foot and soft parts only), Lamelli- branchiata and insect larvae, e.g. Povilla adusta (Ephemeroptera) and Trichoptera 262 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES (including the cases). Each fish had also ingested fairly large quantities of bottom debris. Breeding. Few data are available ; one female 75 mm. S.L. is sexually mature whilst another, 88 mm. S.L. appears to be a juvenile maturing for the first time. Diagnosis and affinities. The nature of its lower pharyngeal bone and dentition places H. theliodon on the same level of structural modification as H. humilior and H. riponianus. In other characters (especially general appearance and oral dentition) H. theliodon is unlike both the former species. It differs from H. humilior in several morphometric characters (straight and not curved dorsal head profile, deeper preorbital, longer snout and smaller eye) and from H. riponianus in the shape of the head and the nature of its oral dentition. Haplochromis theliodon also differs from both the other species in having much smaller and more deeply embedded pectoral scales. The affinities of H. theliodon are not easily determined except in so far as the species is clearly a little-modified derivative of the generalized stem. The small pectoral scales of H. theliodon are not common in generalized species still extant in Lake Victoria ; but, small chest scales do characterize certain of the generalized and fluviatile species of East Africa, some of which also show an incipient hypertrophy of the pharyngeal bones and dentition (Greenwood, unpublished). Study material and distribution records Museum and Reg. No. Locality Collector Tanganyika B.M. (N.H.). 1959.4.28.163 . Majita . E.A.F.R.O. (Holotype) 1959-4- 28. 164-168 . . Uganda 1959.4.28.169 . Jinja Pier . Haplochromis empodisma sp. nov. (Text-figs. 16 and 17) Haplochromis ishmaeli (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 446 ; Idem, 1915, Cat. Afr. Fish. 3, 293 ; Regan, 1922, Proc. zool. Soc. London, 169. Tilapia lacrimosa (part), Boulenger 1906, op. cit., 450 ; Idem, 1915, torn, cit., 234. Haplochromis cinereus (part) Regan, 1922, op. cit., 166. Note. Because two specimens of H . empodisma (collected during Graham's 1927- 28 survey of Lake Victoria) were mistaken for syntypes of H. michaeli Trewavas, this species was referred to as H. michaeli in previous papers (Greenwood, 1954, 1956^). All references to " H. michaeli" in these publications should now be corrected to read Haplochromis empodisma. Holotype. An adult male 117 + 23 mm. total length (B.M. (N.H.) Reg. No. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 263 1959.4.28.170), caught on the bottom in 90 feet of water off the southern tip of Kibibi Island (o 10' N. ; 33 10' E.), Uganda. Description, based on thirty-nine fishes (including the holotype) 65-117 mm. S.L. Depth of body 33-0-43-8 (M = 39-3) per cent of standard length, length of head 33-3-39-4 (M = 36-7) per cent. Dorsal head profile straight or gently curved, sloping at an angle of 35-4o. Preorbital depth 15-1-20-5 (M = 18-1) per cent of head ; least interorbital width 20-6-28-6 (M = 24-3), length of snout 27-5-37-2 (M = 32-9), diameter of eye 24-4-34-0 (M = 27-3) per cent. Depth of cheek shows a marked positive allometry with standard length, being 21-2-23-0 per cent of head in fishes less than 80 mm. S.L. (N = 4) and 23-5-31-4 (M = 27-9) per cent in larger specimens (N = 35). Caudal peduncle 14-5-20-0 per cent of standard length and 1-3-1-7 (mode 1-4) times as long as deep. Mouth horizontal or slightly oblique ; lips not noticeably thickened. Posterior tip of the maxilla reaching the vertical through the anterior margin of the orbit or to FIG. 16. Haplochromis empodisma ; holotype. Drawn by John Norris Wood. below the anterior part of the eye ; less frequently, not reaching as far as the orbit. Lower jaw 39- 1-48-7 (M = 43-9) per cent of head, 1-5-2-3 (mode, ill-defined, 2-0) times as long as broad. Malformation of the lower jaw is relatively common in this species ; individuals so affected have the lower jaw broader than the upper (which closes within the lower) and are distinctly prognathous. Gill rakers usually slender, although relatively stout rakers also occur ; 7-10 (modal range 8-9) on the lower part of the first gill arch, the lowermost 1-3 rakers reduced, occasionally the uppermost I or 2 somewhat flattened and bi- or trifid. Scales ctenoid ; lateral line with 30 (f.io), 31 (f.g), 32 (f.i5) or 33 (f.4) scales. Cheek with 3 or 4 (occasionally 5) series. Six or 7 scales between the lateral line and the dorsal fin origin, 6-8 between the pectoral and pelvic fin bases. Fins. Dorsal with 23 (f.8), 24 (f-3o) or 25 (f.i) rays, anal with 10 (f.i), n (f.22) or 12 (f.i6), comprising XIV-XVI, 8 or 9 and III, 7-9 spinous and branched rays for 264 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES the fins respectively. Pectoral fin 73-0-96-5 (M = 85-8) per cent of head. Caudal truncate or sub-truncate. Teeth. The teeth in the outer row of both jaws are slender and often gently recurved ; in fishes less than 95 mm. S.L. the teeth are unequally bicuspid but in larger individuals are weakly bicuspid or unicuspid ; it is usual to find both types of teeth in large fishes. There are 54-82 (modal range 70-72) teeth in the upper jaw. Teeth forming the inner series are small and usually tricuspid but some unicuspids may occur, particularly in fishes over 100 mm. S.L. ; there are 2 or 3 (rarely 4) rows in the upper jaw and 1-3 in the lower. The innermost row, especially in the upper jaw, is implanted obliquely. A common abnormality affecting both inner and outer teeth is for the crowns to be coarse, slightly swollen and darkly pigmented ; in such teeth the crown is globose. Lower pharyngeal bone and teeth. The lower pharyngeal is fine and rather narrow, o-2cm FIG. 17. Haplochromis empodisma ; lower pharyngeal bone, occlusal 0-2 cm FIG. 1 8. Haplochromis obtusidens ; lower pharyngeal bone, occlusal view. the dentigerous surface having the outline of an isosceles triangle. Usually, none of the pharyngeal teeth is markedly enlarged, but in certain specimens of all sizes a few teeth in the two median rows are slightly coarser. Even when enlarged these teeth retain the same form as their lateral and more slender congeners. Osteology. The neurocranium and dentary of H. empodisma were figured and briefly described in an earlier paper (Greenwood, 1956^, p. 305, fig. 5 ; the species was then wrongly identified as H. michaeli]. Both the neurocranium and the jaw elements are directly comparable with those of a generalized species such as H. macrops, except that in H. empodisma the outer teeth are more numerous. Coloration in life. Sexually active males. Ground colour dark turquoise on the flanks and dorsal body surface, silver-yellow ventrally ; snout and dorsal head surface dark red. Dorsal and caudal fins diffuse red ; anal black anteriorly, dark red posteriorly, the ocelli orange-red. Pelvic fins black. Quiescent males as for females (see below) but dorsal fin with red maculae between the branched rays. Ripening males show some reddening of the head and snout, whilst the maculae of the dorsal A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 265 are more intense. Females. Ground colour yellowish-silver, darker dorsally. Dorsal and caudal fins dark neutral ; anal yellow ; pelvics very pale yellow. Preserved material. Males (sexually active), dark grey becoming black ventrally ; a distinct lachrymal stripe and two bars across the snout (quiescent males greyish brown, dusky on the belly ; very faint indications of up to five incomplete transverse bars on the flanks ; lachrymal stripe visible but no bars across the snout). Dorsal fin dark, the lappets black ; caudal dark proximally, lighter distally ; basal third of anal black, remainder of fin light, with two or three large white ocelli. Pelvic fins black. Females brownish-silver ; in a few specimens there are very faint indications of a lachrymal stripe and of incomplete transverse bars on the flanks. All fins hyaline ; in some fishes the upper half of the caudal is slightly maculate. Distribution. Known only from Lake Victoria. Ecology. Habitat. It seems that H. empodisma is generally restricted to those areas of the lake where the bottom is composed of organic mud. The species has been caught in bottom nets set at depths from 10-90 feet. Food. The stomach and intestinal contents of thirty -seven fishes (from numerous localities) show that H. empodisma is a bottom feeder. The principal food organisms are the larval (and less frequently pupal) stages of dipterous insects, together with diatoms derived from the bottom mud. Only two fishes yielded remains of Mollusca ; in one fish a few fragments of bivalve shell were found and in the other an operculum from a large gastropod. Since no other snail fragments were found in this fish, it is possible that the operculum was accidentally ingested. Breeding. It is not known whether H. empodisma spawns in the habitats described above or whether spawning takes place over a more solid substrate, for example outcrops of rock or sand. Females carrying embryos and larvae have been caught together with non-breeding fishes. The smallest sexually active fish is a female 84 mm. S.L. ; the smallest adult male is 90 mm. S.L. Both sexes reach the same maximum adult size. Affinities. Structurally and in its feeding habits, H. empodisma must be considered a generalized species. It differs from the majority of generalized forms in Lake Victoria only by its larger size, greater number of teeth (possibly a correlate of the larger size) and in the wide range of depths at which it has been caught. The nearest living relative of H. empodisma is a small and as yet undescribed species which occurs in the same habitat but is confined to shallow water. Haplochromis empodisma and H. obtusidens are strikingly similar except for one structure, the lower . pharyngeal bone. In H. obtusidens the lower pharyngeal is thickened and carries a number of enlarged, crushing teeth, whereas in H. empodisma this bone is slender and carries numerous fine teeth. Haplochromis empodisma could well represent the ancestral type from which H. obtusidens was derived by an increase in size of the pharyngeal bones and a correlated change in the pharyngeal dentition. Although the inter-specific differences in the nature of the pharyngeal mill are fairly clear-cut, some individuals deviate from the specific mode in such a way as to indicate a likely transitional condition in the evolution of an " obtusidens "-like species. 266 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Since, in turn, H. obtusidens provides a structural type basic to the evolution of a specialized crushing pharyngeal mill (as found in H. ishmaeli and H. pharyngo- mylus), H. empodisma could represent the extant representative of the basal species in the phyletic line which culminated in H. ishmaeli and H , pharyngomylus. Study material and distribution records Museum and Reg. No. Locality Collector Uganda B.M. (N.H.). 1959.4.28.170 . Off Kibibi Island . E.A.F.R.O. (Holotype) ,, 1906.5.30.472 . Entebbe . Degen ,, 1966.5.30.402 . Bunjako . ,, (Paratype, H. ishmaeli) 1906.5.30.404 (Paratype, H. ishmaeli) 1959.4.28.180-189 . Pilkington Bay . E.A.F.R.O. ,, 1959.4.28.194-195 . Thruston Bay . ,, 1959.4.28.196-197 . Ekunu Bay . ,, 1959.4.28.198 Jinja 1959.4.28.199 . Napoleon Gulf, near Jinja . ,, ,, 1959.4.28.202-203 . Kibibi Island . ,, 1959.4.28.358 . Off S. tip of Kibibi Island . ,, Tanganyika ,, 1959.4.28.190-193 . Nyamakyamwa . E.A.F.R.O. Kenya 1959.4.28.200-201 . Manadu Island . ,, ,, 1959.4.28.171-179 . Off Port Southby . ,, Lake Victoria, Locality Unknown 1928.6.2.37-38 . . M. Graham Haplochromis obtusidens Trewavas 1928 (Text-fig. 18) Haplochromis desfontainesi (part), Boulenger, 1915, Cat. Afr. Fish., 3, 302. Tilapia lacrimosa (part), Boulenger, 1915, torn, cit., 234. Haplochromis cinereus, (part) Regan, 1922, Proc. zool. Soc. London, 166. Haplochromis obtusidens Trewavas, 1928, Ann. Mag. nat. Hist. (10), 2, 95. Lectotype. An adult male 107-0 -f 23-0 mm. total length (B.M. (N.H.) Reg. No. 1928.5.30.21). Description, based on forty-four specimens (including the lectotype and one paratype) 60-114 mm. S.L. Certain proportions show fairly well-marked allometry with standard length ; for these, ranges and means are given for each of the relevant size-groups. Depth of body 35-0-44-3 (M = 38-6) per cent of standard length, length of head 32-5-38-2 (M = 35-8) per cent. Dorsal head profile gently curved or, less commonly, straight, sloping at an angle of 4o-5o. Preorbital depth in fishes < 70 mm. S.L. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 267 (N = 4), 13-0-16-7 (M = 14-5) per cent of head, in larger fishes (N = 40) 15-1- 20-5 (M = 17-9) ; least interorbital width 21-8-29-0 (M = 24-2) per cent ; snout length in fishes < 85 mm. S.L. (N = 6), 26-0-31-0 (M = 28-5) per cent, in larger individuals (N = 38), 29-0-36-4 (M = 33-3) ; diameter of eye in fishes < 85 mm., 27-6-34-8 (M = 31-2) per cent and in larger fishes 24-3-30-8 (M = 27-2) ; depth of cheek in specimens < 85 mm., 18-6-23-8 (M 21-5) and in larger individuals 21-2- 30-0 (M = 26-7) per cent of head. Caudal peduncle 15-2-19-7 (M = 17-5) per cent of standard length, 1-2-1-8 (mode 1-5) times as long as deep. Jaws equal anteriorly ; mouth horizontal or slightly oblique ; lips not markedly thickened. Posterior tip of the maxilla extending to the vertical through the anterior orbital margin or to below the anterior part of the eye ; exceptionally, not reaching the orbit. Lower jaw 37-9-45-5 (M = 41-8) per cent of head and 1-3-2-0 (mode 1-6) times as long as broad. Gill rakers relatively slender, although short and stout rakers are found in a few individuals ; 7-9 (modal range 7-8), rarely 6, on the lower part of the first arch, the upper pair of rakers often flattened. Scales ctenoid ; 29 (f.i), 30 (f.7), 31 (f.2i), 32 (f.n) or 33 (f.4) in the lateral line ; cheek with 3 or 4 (rarely 5) series ; 5^-7 scales between the lateral line and the dorsal fin origin, 5-7 (rarely 9) between the pectoral and pelvic fin bases. Fins. Dorsal with 23 (f.i3), 24 (.25) or 25 (f.6) rays, anal with n (.24), 12 (f.ig) or 13 (f.i), comprising XIV-XVI, 8-10 and III, 8-10 spinous and branched rays for the fins respectively. Pectoral 73-5-103-0 (M = 86-8) per cent of head. Caudal truncate or sub-truncate. Teeth. In fishes less than 90 mm. S.L., the outer row in both jaws is composed of unequally bicuspid and relatively slender teeth. Larger individuals have an ad- mixture of either unequally bicuspid and unicuspid teeth or of weakly bi- and unicuspids. A few individuals have exclusively bicuspid or unicuspid teeth. There are 40-80 teeth (ill-defined mode at 70 ; modal range 66-70) in the outer row of the upper jaw. Teeth in the inner series are generally tricuspid and small ; in some fishes over 100 mm. S.L. there may be either a mixture of tri- and unicuspid teeth or, more rarely, only unicuspids. The innermost row of teeth (especially in the upper jaw) is implanted obliquely and in some individuals the whole inner series lie at an oblique angle. There are 2 or 3 (rarely i or 4) rows of inner teeth in the upper jaw and I to 3 (usually 2, rarely 4) in the lower. Lower pharyngeal bone and teeth. Despite some individual variability, the lower pharyngeal bone of H. obtusidens is always obviously thickened and the two median rows of teeth are enlarged and molariform or sub-molariform. As might be expected, there is a positive correlation between size and the degree to which the pharyngeal bones and teeth are enlarged. In fishes over 90 mm. S.L. as many as six rows of teeth may be composed of molariform elements ; even in specimens with a few molariform rows the remaining teeth are enlarged, except in the upper corners of the bone. Osteology. In all respects except the form of the pharyngeal apophysis, the neuro- cranium of H. obtusidens resembles that of H . empodisma ; the pharyngeal apophysis, 268 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES however, is stouter and broader, thereby foreshadowing the condition found in H. ishmaeli and H. pharyngomylus. In both these species the basioccipital facets are more expanded than in H. obtusidens. Coloration in life. Sexually active males. Ground colour light blue-black with, ventrally, a silver patch extending from the isthmus almost to the vent ; branchio- stegal membrane black. Dorsal fin dark with red lappets and margin to the soft part and red spots or dashes between the branched rays. Caudal dark with a blood- red margin and a diffuse red centre. Anal dark proximally, blood-red distally ; ocelli orange or orange-red. Pelvic fins black. Quiescent males as for females but lacking the red spots or flush on the dorsal fin. Females. Ground colour silver-grey. Dorsal fin dark with traces of a red flush. Caudal dark. Anal and pelvic fins yellowish. Preserved material. Adult males. Ground colour brownish, dusky ventrally (below the level of the lower lateral line), with faint indications of four transverse stripes originating from the dark area but not reaching the base of the dorsal fin ; the intensity of the dusky area varies considerably, from charcoal to coal-black. Dorsal fin hyaline to dusky, the basal region and the lappets black ; soft part of the fin sometimes maculate. Caudal dark on its proximal half, light (yellowish) distally. Anal dark on the basal half to third, light distally, with two large white ocelli. Pelvic fins black. Females silver-grey to yellowish-brown ; in some fishes there are very faint indications of four to six, incomplete transverse bars on the flanks. Dorsal fin dark or hyaline. Caudal dark, the upper half maculate in some individuals. All other fins hyaline. Note on Jour atypical individuals. Four specimens, 80-94 mm. S.L. (B.M. (N.H.) Reg. Nos. 1959.4.28.236-239) from Old Bukakata Bay, Uganda are included in the description given above although they differ from the generality of specimens in certain characters. The lower pharyngeal bones in two of these fishes are somewhat more slender than is modal but are typical in the other two specimens. All four fishes have fewer upper outer teeth than is usual (40-52 cf. modal range 66-70) and somewhat shallower cheek than equivalent sized specimens from other areas. In this latter character, however, they resemble the two type specimens. Finally it must be mentioned that these fishes resemble one another in general facies rather more closely than they resemble the other specimens. Such obvious but undefinable and geographically localized facies are fairly common amongst Lake Victoria Haplochromis species. The exact status of the four specimens from Bukakata cannot be determined for want of more material from this locality. Distribution. Known only from Lake Victoria. Ecology. Habitat. Haplochromis obtusidens is predominantly a species of shallow water (less than 30 feet), apparently restricted to a substrate of soft, organic mud ; a few individuals have, however, been taken over sand and in water about 60 feet deep. Food. The gut contents of forty-six fishes (from numerous localities) indicate that the principal food organisms of H. obtusidens are insects (especially larval Diptera) and molluscs (particularly the bivalve Corbicula, although some gastropods A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 269 [Melanoides] are also eaten). Together with these organisms, the fishes had ingested fairly substantial quantities of bottom mud, which, in the areas inhabited by H. obtusidens, is almost entirely composed of living and moribund diatoms. A com- parison of stomach and intestinal contents shows that the protoplasm of these diatoms is digested by the fish. Other plants, in this case mostly blue-green algae, are apparently undigested. Breeding. Haplochromis obtusidens is a female mouth-brooder ; exact spawning sites were not discovered. A male 83 mm. S.L. is the smallest mature fish in the sample studied ; the smallest mature female is 89 mm. S.L. Both sexes reach the same maximum adult size. Affinities. The relationships of H. obtusidens seem to lie with H. empodisma which it resembles in body-form and most syncranial characters. Haplochromis obtusidens differs in having an enlarged neurocranial apophysis for the upper pharyngeal bones and in the correlated character, enlarged pharyngeal bones and teeth. The species also shows certain fairly marked affinities with H. ishmaeli and H. pharyngomylus and could well represent the extant version of an ancestral type from which these two species evolved (see also p. 265). In Lake Edward, H. malacophagus Poll has reached a comparable evolutionary stage leading towards extreme pharyngeal hypertrophy, but the two species are not closely related. Study material and distribution records Museum and Reg. No. B.M. (N.H.). 1906.5.30.531 ,, 1959.4.28.204-212 ,, 1959.4.28.213-218 1959.4.28.228-229 1959.4.28.236-239 . 1959.4.28.243-244 . 1959.4.28.245 1959.4.28.247 Locality Uganda Buganga Pilkington Bay Ekunu Bay Pilkington Bay Old Bukakata Bay Dagusi Island Buka Bay Sesse Islands Collector Degen E.A.F.R.O. 1909.7.27.43 1959.4.28.219-227 1959.4.28.230-232 J 959 .4.28. 240-241 1959.4.28.242 1959.4.28.246 1959.4.28.248 Kenya Kavirondo Bay Off Port Southby Kach Bay (Kavirondo Gulf) Beach below Usoma lighthouse Off Port Southby Kavirondo Gulf Off mouth of Nzoia River Alluaud E.A.F.R.O. Tanganyika ,, 1959.4.28.233-235 . Beach near Majita Lake Victoria, Locality Unknown 1928.5.30.21 (Lectotype) B.M. (N.H.). 1928.5.20.20 (Paratype) M. Graham 270 A REVISION OF THE LAKE VICTORIA HA PLOCHROMIS SPECIES Haplochromis pharyngomylus Regan 1929 (Text-figs. 19 and 20) Haplochromis ishmaeli (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 446; Idem, 1915, Cat. Afr. Fish., 3, 293; Regan, 1922, Proc. zool. Soc. London, 169. Haplochromis pharyngomylus Regan, 1929, Ann. Mag. nat. Hist. (10) 3, 388. Description, based on thirty-eight specimens (including the holotype) 70-126 mm. S.L. Depth of body 33-8-42-0 (M = 38-5) per cent of standard length, length of head 31-5-36-8 (M = 34-6) per cent. Dorsal head profile straight or somewhat curved (occasionally concave between the eyes), sloping fairly steeply. Preorbital depth 13-8-19-0 (M = 16-8) per cent of head, least interorbital width 23-7-28-5 (M = 26-3), length of snout 27-3-33-3 (M = 30-8), diameter of eye 23-0-31-8 (M = 26-5), depth of cheek 19-7-27-0 (M = 24-1) per cent of head. Caudal peduncle 13-6-18-5 (M = 16-2) per cent of standard length, 1-1-1-6 (modal range 1-3-1-5) times as long as broad. Mouth horizontal or slightly oblique, jaws equal anteriorly, the lower 35-8-44-0 (M = 38-6) per cent of head, 1-3-2-0 (mode 1-4) times as long as broad. Posterior tip of the maxilla reaching or almost reaching the vertical to the anterior orbital margin. Gill rakers moderately stout, 6-8 (mode 7), rarely 9, on the lower part of the first arch, the lowermost i or 2 rakers reduced. Scales ctenoid ; lateral line with 30 (f.i), 31 (f.6), 32 (1.12), 33 (1.13), 34 (1.4) or 35 (f.i) scales ; cheek with 3 or 4 series. Six or 7 (less frequently 8) scales between the lateral line and the dorsal fin, 7 or 8 (less frequently 6 or 9, rarely 10) between the pectoral and pelvic fin bases. Fins. Dorsal with 24 (f.7), 25 (f.28) or 26 (f.3) rays, anal with n (f.5), 12 (f.28) or 13 (f-5), comprising XV-XVI, 8-10 (rarely n) and III, 8-10 spinous and branched rays for the fins respectively. Pectoral 68-5-91-0 (M = 79-6) per cent of head. Caudal truncate. Teeth. In fishes 70-90 mm. S.L. only unequally bicuspid, relatively stout teeth occur in the outer row of both jaws. Fishes 90-100 mm. S.L. show some variability in the form of these teeth, which may be unicuspids, a mixture of bi- and weakly bicuspid teeth or both bi- and unicuspids. There are 30-42 (mode 36, modal range 36-40) outer teeth in the upper jaw. Teeth in the inner rows are usually tricuspid in fishes less than 100 mm. S.L. In larger individuals these teeth may be tricuspid or there can be a mixture of tri- and weakly tricuspids ; it is uncommon to find only unicuspid teeth in the inner rows. There are 2 or 3 (rarely 4) rows in the upper jaw and i or 2 (rarely 3) in the lower. Lower pharyngeal bone and teeth. With one exception, the lower pharyngeal bone is massive, even in the smallest specimens. Nevertheless, a slight size correlated increase in relative stoutness can be detected in large fishes. Apart from H. ishmaeli (p. 277) and Astatoreochromis alluaudi, no other Lake Victoria Cichlidae have such large pharyngeal bones. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 271 The exceptional specimen mentioned above was caught near Jinja and measured 98 mm. standard length. The lower pharyngeal of this fish can be compared with that of an H. obtusidens of a similar size. Since in all other respects this fish resembles H. pharyngomylus more closely than H. obtusidens it is included in the description of H. pharyngomylus. Four other specimens with even finer pharyngeal bones and dentition, but otherwise resembling H. pharyngomylus, are not included in the description but are dealt with in a separate appendix (p. 274). The lower pharyngeal teeth of H. pharyngomylus are large and molariform except for a few teeth situated in the upper, lateral angles of the bone. These teeth, although relatively stout, are cuspidate and small ; the number of such teeth decreases markedly in the largest individuals. By analogy with other species having FIG. 19. Haplochromis pharyngomylus ; holotype. Drawn by Miss M. Fasken. enlarged pharyngeals and for which a greater size range of specimens is available is seems likely that small H. pharyngomylus should have less massive bones and fewer molariform teeth, confined to the median tooth series (see Greenwood, 1959). Osteology of the neurocranium. Apart from an enlarged and strengthened articular apophysis for the upper pharyngeals, the neurocranium of H. pharyngomylus is that of a large, generalized Haplochromis. The apophysis has been figured and described previously (Greenwood, 1954). Compared with the apophysis of H. obtusidens that of H. pharyngomylus has a greater surface area and the prootic buttresses are more obvious. These characters are clearly correlated with the more massive pharyngeal bones and musculature of H. pharyngomylus. Only slight dif- rences exist in the apophyseal region of H. pharyngomylus and H. ishmaeli ; but both species differ from Astatoreochromis alluaudi in the form taken by the various elements contributing to the apophysis (Greenwood, 1959. In that paper, I also 272 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES briefly described the possible effects and interactions of genetical and environmental factors on the development of apophyseal form and size in mollusc-crushing species). Coloration in live fishes. Adult males. Ground colour blue-grey overlying silver ; a distinct coppery sheen on the flanks. Dorsal fin hyaline with pinkish lappets and margin to the soft part. Anal hyaline, ocelli yellow. Caudal hyaline with a pink flush most intense distally and on the ventral half of the fin. Pel vies black. Females. Golden-green, becoming silvery- white ventrally. Dorsal and anal fins dark ; the caudal yellowish-green, darker proximally. Preserved material. Adult males. Greyish, darker in sexually active individuals ; a dark lachrymal stripe sometimes visible ; barring on the flanks variable, but usually consisting of seven to ten narrow vertical bars and a very faint longitudinal stripe following the course of the upper lateral line. Dorsal fin hyaline, dark basally in juvenile and quiescent fishes but almost black in sexually active individuals ; lappets dark, the soft dorsal intensely maculate in sexually active fishes. Anal hyaline, dark basally, with five or six ocelli arranged in one or two horizontal rows. o-2cm\T7 0-2 cm (a) (b) FIG. 20. Haplochromis pharyngomylus ; lower pharyngeal bone, in occlusal view, of (a) a fish 98 mm. S.L., and (b) a fish 126 mm. S.L. Caudal hyaline in quiescent fishes, dusky in active ones. Pelvics hyaline in juveniles, black in adults. Females, silvery-grey to light brown ; striping variable ; when present, as described for males. Dorsal fin hyaline or greyish, the soft part weakly maculate in adults. Caudal hyaline or dark, the upper half faintly maculate in some specimens. Anal and pelvic fins hyaline. Distribution. Known only from Lake Victoria. Ecology. Habitat. Haplochromis pharyngomylus is apparently confined to water less than 40 feet deep and to areas of the lake where the bottom is hard (sand, shingle or, less frequently, rock) ; the species is often found amongst stands of aquatic plants. A few specimens have been caught in the areas of intergradation between sand and mud substrates. Food. Analyses of stomach and intestinal contents from thirty specimens (repre- senting most localities) show that fishes in the size-range 75-115 mm. S.L. feed mainly or even exclusively on Mollusca ; only one fish had eaten mollusca and insect larvae. The sample examined also indicated that both bivalves and gastropods A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 273 are eaten in approximately equal proportions. The fragmented shells do not permit accurate identification of the prey species ; Sphaerium sp. and Corbicula sp. were recognized amongst the bivalve remains and Melanoides tuberculata (the pre- dominant snail), Bellamy a and less frequently Biomphalaria amongst the gastropod fragments. It is clear from the small size of most shell fragments that the pharyngeal mill in H. pharyngomylus is an efficient crushing mechanism. Yet, despite this powerful barrier some shells do pass into the stomach almost undamaged. Since these shells are invariably empty it seems that the digestive enzymes (especially those of the stomach) are capable of breaking down the bodies of snails without preliminary and physical assistance from the pharyngeal teeth. Trewavas (1938) noticed that the gut contents of H. mahagiensis (a mollusc eater from Lake Albert) were " im- pregnated and held in a hard mass by botryoidal aggregates of calcite. Whether these were formed before or after death and fixation is a matter for conjecture." My observations on the gut contents of H. pharyngomylus support Trewavas' observations only when the material had been fixed in formol. This would suggest that the calcite aggregates are formed as a result of a chemical reaction between the slightly acid formol and the calcium of the shell. I have never observed aggre- gates in fresh gut contents, where the various shell fragments could be separated easily. Breeding. The actual spawning sites and breeding behaviour of H. pharyngomylus are unknown, but females carrying embryos and larvae in the buccal cavity occur together with non-breeding fishes in most localities. The smallest adult recorded is a female 90 mm. S.L. ; the smallest adult male is 94 mm. S.L. Both sexes reach the same maximum adult size. Affinities and diagnosis. Extreme hypertrophy of the pharyngeal mill serves to set H. pharyngomylus apart from all except one species of Lake Victoria Haplo- chromis. The other species is H. ishmaeli which appears to be very closely allied to H. pharyngomylus. Although in structural characters the two species are similar an experienced observer can, in most instances, readily distinguish between the two species. But, as is so often the case with Haplochromis, the subjective characters used for " field " identifications cannot be quantified or adequately described on paper. When seen alive, adult males of the two species have distinctive coloration. Haplochromis pharyngomylus is distinguished from H . ishmaeli by the following characters : fewer outer teeth in the upper jaw (30-42, modal range 36-40 cf. 38-66, modal range 44-52 in H. ishmaeli); shorter pectoral fin (68-0-91-0, M = 79-6 per cent of head cf. 75-0-102-0, M = 88-5) ; differences in male breeding coloration (see p. 272 and p. 277) and differences in habitat preference (H. pharyngomylus is es- sentially a species of hard substrates whilst H. ishmaeli shows a marked preference for muddy areas) . With the exception of differences in male breeding colours, none of these characters alone is trenchant ; taken together, however, they provide fairly reliable diagnostic features. The phylogenetic position of H. pharyngomylus has been discussed above (p. 269) and in an earlier paper (Greenwood, 1954) . In Lake Edward, H. placodus Poll represents the equivalent evolutionary phase in 274 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES the development of hypertrophied pharyngeals. Haplochromis placodus and H. pharyngomylm could be derived from a common ancestral stem, but of course, the two species could equally well be examples of convergent evolution. APPENDIX Four specimens (115-120 mm. S.L.) from Jinja (B.M. (N.H.) Reg. No. 1959. 4.28.352-355) bear a strong resemblance to H. pharyngomylus except that the lower pharyngeal bone in these specimens is barely enlarged and the pharyngeal dentition is correspondingly weak. These fishes cannot be distinguished from H. pharyn- gomylus on proportional measurements, teeth of jaws, fin and scale counts, general appearance or preserved coloration. All four fishes are adult males but unfortunately their live coloration was not recorded. From H. pharyngomylus material described above and from field observations on numerous other specimens it is clear that intraspecific variability in the enlargement of the pharyngeal bones is slight and mainly correlated with size. No specimens were found which could be considered intermediate between the typical condition for the species and that seen in the aberrant individuals. On the other hand, a study of Astatoreochromis alluaudi (a Haplochromis-like monotypic genus) showed that some populations have enlarged pharyngeals whilst others exhibit only slight hypertrophy of these bones (Greenwood, 1959). Thus for the moment it is impossible to dismiss the possibility that the four " pharyngomylus "-like fishes are indeed aberrant members of that species. I do not propose, however, to include them in this revised description or to describe them as distinct species until further collections dictate one step or the other. Study material and distribution records Museum and Reg. No. Locality Collector Uganda B.M. (N.H.). 1906.5.30.383 . Entebbe . Degen 1 907-5- 7-7i Buddu Coast . Simon 1958.12.5.30-33 . Entebbe . Pitman 1959.4.28.317-321 . Beach near Nasu Point E.A.F.R.O. 1959-4-28.322-326 . Entebbe, Harbour 1959.4.28.327-329 . Entebbe, Airport beach . ,, 1 959- 4- 28. 333-334 - Beach near Grant Bay . ,, 1959.4.28.335-336 . Between Yempita and . E.A.F.R.O. Busiri Isls., Buvuma Channel 1959-4-28.340-343 . Beach near Hannington Bay 1959.4.28.344 . Entebbe, Harbour 1959.4.28.345-347 . Napoleon Gulf, near Jinja 1959-4-28.351 . Ramafuta Island 1959.4.28.358 ". Jinja Pier Kenya 1959.4-28.348 . Kasingiri Gingo (Kavirondo Gulf) A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 275 \ Study material and distribution records (cont.) Museum and Reg. No. Locality Collector Tanganyika 1959.4.28.337-339 . Majita ,, 1959.4.28.349 . Beach near Majita . ,, ,, 1959.4.28.350 . Mwanza, Capri Bay . ,, Lake Victoria, Locality Unknown 1928.5.24.313 . . M. Graham (Holotype) 1959.4.28.330-332 . E.A.F.R.O. Haplochromis ishmaeli Blgr. 1906 (Text-fig. 21) Haplochromis ishmaeli (part) Boulenger, 1906, Ann. Mag. nat. Hist. (7), 17, 446 ; Idem, 1915. Cat. Afr. Fish. 3, 293 ; Regan, 1922, Proc. zool. Soc. London, 169. Tilapia pallida (part) Boulenger, 1911, Ann. Mus. Genova (3), 5, 74; Idem, 1915, torn, cit., 231-2. Labrochromis pallidus Regan, 1920, Ann. Mag. nat. Hist. (9), 5, 45 (footnote). Tilapia martini (Part) Boulenger, 1915, torn, cit., 239. Paratilapia victoriana (part) Boulenger, 1915, torn, cit., 341. Haplochromis macrops (part) Regan, 1922, Proc. zool. Soc. London, 166. In the original description of H. ishmaeli, Boulenger listed thirteen type specimens (syntypes) all collected by Degen at Bunjako, Uganda. However, only eleven specimens (ten in spirit and one skeleton) answering this description can be found in the collections of the B.M. (N.H.) ; furthermore, only eleven such specimens are recorded in the Museum's catalogue of accessions. That the number thirteen was a slip of the pen seems certain because Boulenger (1915) only lists ten types from Bunjako in the third volume of his Catalogue of Fresh- water Fishes of Africa. The only other specimen from this locality is listed in the " Catalogue as a skeleton and is presumably the eleventh syntype. These figures agree with the number of specimens I could locate and, more im- portant, they agree with the number registered in the Museum's record of accessions. But, the situation is still somewhat obscure because in the " Catalogue " (torn. cit. loc. cit.} Boulenger lists as a type a specimen collected from Entebbe ; since no reference is made to this fish in the original description I am treating its later inclusion in the list of types as erroneous. Note on the genus Labrochromis, Regan, 1920. This genus, briefly described in a footnote to Regan's paper on the genera of Tanganyika cichlids was apparently based on a single specimen prepared from one of the syntypes of T. pallida (B.M. (N.H.) Reg. No. 1911.3.3.132). In his revision of the Lake Victoria Cichlidae, Regan (1922) correctly identified this skeleton as being that of Haplochromis ishmaeli and abandoned the genus Labrochromis on the grounds that, apart from possessing hypertrophied pharyngeals, H. ishmaeli is " nearly identical with H. cinereus." Whilst I do not agree with the latter part of this statement, I fully endorse Regan's action in not maintaining the monotypic genus Labrochromis for Haplochromis ishmaeli. ZOOL. 6, 4 19 276 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES Lectotype of Haplochromis ishmaeli. An adult female 104-0 + 23-0 mm. total length (B.M. (N.H.) Reg. No. 1906.5.30.400) collected by Degen at Bunjako, Uganda. Description, based on thirty-five specimens (including the lectotype and two paratypes) 82-136 mm. S.L., all from Lake Victoria. Two specimens from Lake Edward are described separately on p. 278. Depth of body 37-0-45-5 (M = 40-1) per cent of standard length, length of head 33'8-37'5 (M = 34-8) per cent. Dorsal head profile slightly curved or straight, sloping fairly steeply. Preorbital depth 15-3-20-5 (M = 17-0) per cent of head, least interorbital width 24-0-32-0 (M = 27-6), length of snout 29-0-36-0 (M = 31-6), diameter of eye 23-0-31-0 (M = 27-7), depth of cheek 20-7-31-0 (M = 25-5) per cent. Caudal peduncle 14-6-18-8 (M 17-6) per cent of standard length, 1-2-1-6 FIG. 21. Haplochromis ishmaeli : lectotype (from Boulenger, Pishes of the Nile). (modal range 1-3-1-5) times as long as deep. Jaws equal anteriorly, the lower 35-8-42-5 (M = 39- 1) per cent of head and 1-4-2-0 (modal range 1-4-1-6) times as long as broad. Mouth horizontal or very slightly oblique ; the posterior tip of the maxilla reaching the vertical through the anterior orbital margin or somewhat beyond, rarely not quite reaching the orbit. Gill rakers stout, but relatively slender in a few specimens ; 6-8 (mode 7), rarely 9, on the lower part of the first arch, the lowermost one or two rakers usually reduced. Scales ctenoid ; lateral line with 30 (f.3), 31 (f.8), 32 (f.i5), 33 (f-5) or 34 (f.4) scales ; cheek with 3 or 4 rows. Five to 7 (rarely 8) scales between the lateral line and the dorsal fin origin ; 8-9 (occasionally 7) between the pectoral and pelvic fin bases. Fins. Dorsal with 23 (f.i), 24 (1.13) or 25 (f.2i) rays, anal with n (1.13), 12 (f.2i) or 13 (f.i) comprising XV-XVI, 8-10 and III, 8-10 spinous and branched rays for the fins respectively. Pectoral fin 75-0-102-0 (M 88-5) per cent of head. Caudal truncate or subtruncate. Teeth. The outer row in both jaws usually contains a mixture of relatively stout, unequally bicuspid and unicuspid teeth ; less frequently only unicuspids occur in A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 277 this row. There is apparently no size-correlated difference in the type or the number of teeth present. There are 38-66 (modal range 44-52) teeth in the outer row of the upper jaw. Teeth in the inner series are generally tricuspid, but in some fishes the entire inner series are composed of unicuspids. There are 2 or 3 (rarely i) rows of teeth in the upper jaw and i or 2 (less commonly 3) rows in the lower. The inner teeth are implanted somewhat obliquely. Lower pharyngeal bone and teeth. The lower pharyngeal bone in H. ishmaeli is massive and almost all the teeth are molariform ; only those situated in the posterior angles of the bone remain small, with pointed crowns. There is some individual variation in the relative enlargement of the bone, but this is slight in comparison with the variation known from such species as H. obtusidens and Astatoreochromis alluaudi. A slight size-correlated increase in the relative stoutness of the pharyngeal bones was observed in the material studied ; likewise, there is an increase in the number of molariform teeth in larger fishes. In size, shape and dentition, the pharyngeal bones of H. ishmaeli are directly comparable with those of H. pharyngomylus. Neurocranial osteology. The neurocranium of H. ishmaeli is virtually identical with that of H. pharyngomylus. Coloration of live fishes. Sexually active males. Ground colour light yellow-green dorsally, shading to yellow on the flanks and greyish-white ventrally. Dorsal fin yellow-green, lappets of the anterior spines dusky, the remainder scarlet, as are the spots and dashes between the branched rays. Caudal fin greyish, with red maculae (sometimes coalesced) between the rays. Anal smoky grey with black lappets and an overall scarlet flush ; ocelli yellow. Pelvic fins black. Females golden-green shading to silvery-white ventrally, the pectoral region faintly blackish. Dorsal fin hyaline, with a narrow red margin. Anal light yellowish-green. Caudal and pelvic fins pale yellow. Preserved material. Adult males, yellowish-grey, dusky on the chest and branchio- stegal membrane (lighter in sexually quiescent fishes) ; very faint indications of six or seven vertical stripes on the flanks. (These bars are more widely spaced than in H. pharyngomylus, see p. 272). A distinct lachrymal stripe. Dorsal fin hyaline in juvenile fishes but darker and with the soft part maculate in mature males ; lappets dark. Caudal hyaline, the upper part often maculate. Anal hyaline, the basal third dark in sexually active individuals ; three to five large, dead-white ocelli arranged in one or, rarely, two rows. Females, brownish to yellowish-silver ; sometimes, very faint traces of five to seven transverse bars on the flanks ; a weak lachrymal stripe visible in some fishes. All fins hyaline, the dorsal darkest. Distribution. Lakes Victoria and Edward. Ecology. The data given in this section relate to fishes from Lake Victoria ; nothing is known for fishes from Lake Edward. Habitat. Haplochromis ishmaeli is essentially a species of inshore regions where the water is less than 30 feet deep and the bottom composed of soft, organic mud. 278 A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES No specimens have been recorded from depths greater than 60 feet, but some have been caught in nets set over sand and shingle substrates. Haplochromis ishmaeli would seem to be the ecological (soft substrate) counterpart of H. pharyngomylus (solid substrates). Because of the difficulty in distinguishing H. ishmaeli from H. pharyngomylus and especially because the latter species was not recognized until after Graham's col- lections were brought to England, it is impossible to use Graham's (1929) catch records as an additional source of information on the intralacustrine distribution of H. ishmaeli. Food. The stomach and intestinal contents of nineteen fishes (from several localities) indicate that H. ishmaeli in the size-range 84-135 mm. S.L. feed almost exclusively on Mollusca ; the few insect larvae found together with snails in the stomach of one fish suggest that insects may be ingested accidentally. The number of fishes with food remains is insufficient to determine whether bivalves or gastropods predominate in the diet. Slightly more gastropods (Melanoides tuberculata) than bivalves (Corbicula sp.) were found in the sample examined. Breeding. No information is available on the breeding sites or habits of H. ishmaeli although the number of females with " spent " ovaries and ventrally distended mouths suggests that the species is probably a female mouth-brooder. The smallest mature females are 97 mm. S.L. (but see Appendix 2) and the smallest adult male is 98 mm. S.L. Females apparently reach a greater maximum size than do males. Affinities and diagnosis of H. ishmaeli are discussed on page 273, with reference to its closest relative, H. pharyngomylus. APPENDIX (i) Haplochromis ishmaeli from Lake Edward. I have been able to examine only two specimens from Lake Edward, one an adult female 120-0 -f 28-0 mm. long and the other an adult male 118-0 + 27-0 mm. long (see Trewavas, 1933). In general appearance, in most morphometric characters and in scale and fin ray counts the two fishes are indistinguishable from specimens of a similar size from Lake Victoria. The nature of the pharyngeal bones and dentition is also identical. Nevertheless, the Lake Edward fishes do differ slightly in three characters. (i) The interorbital is somewhat narrower (24-1 and 25-0 per cent of head) than the mean interorbital width of Victoria fishes (27-6 per cent) although still within the range known from this population, (ii) The number of outer teeth in the upper jaw (40 and 42) is in the lower section of the range for Lake Victoria fishes, (iii) The caudal peduncle is stouter in the majority of Victoria specimens. Now that more specimens are available from Lake Victoria, Trewavas' (op. cit.) observation on the larger eye of the Edward specimens is no longer applicable. None of the differences commented upon above is so marked as those charac- terizing the Lake Victoria and Lake Edward populations of Astatoreochromis alluaudi (Greenwood, 1959). Nothing is known about the bionomics of H. ishmaeli in Lake Edward ; the two specimens studied were caught in a seine net fished from the eastern shore of the lake near Kisenyi. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 279 (2) Three H. ishmaeli-\ike fishes from Lake Victoria. Three specimens (82, 96 and 100 mm. S.L. (B.M. (N.H.) Reg. Nos. 1959.4.28.304-306)) caught in a trawl off the mouth of the Nzoia river, Kenya, present something of a problem. In appearance and in all standard counts and measurements these fishes are typically H. ishmadi. However, their collector, Mr. S. H. Deathe, recorded the live coloration of these fishes as bright pink. Because all three specimens are adult males, I consider that this striking departure from the usual male coloration of H. ishmaeli may be significant, especially since pink is not one of the basic colours in H. ishmaeli pigmentation. A further interesting difference is that the smallest specimen is sexually mature whereas the smallest adult H. ishmaeli recorded is 98 mm. S.L. Unfortunately I have only these three specimens and I did not see them when alive. Thus, it is difficult for me to assess fully this seemingly outstanding difference in coloration. No typical H . ishmaeli were reported in the same haul and I do not consider that there is enough evidence to decide whether these peculiar individuals represent an aberrant population of H. ishmaeli or a distinct species differing from H. ishmaeli in the coloration of its adult males. Study material and distribution records Museum and Reg. No. B.M. (N.H.). 1906.5.30.400 (Lectotype) ,, 1906.5.30.401-4020 (Paratypes) 1906.5.30.396 (Paratype) 1911.3.3.131 (Paratype Tilapia pallida) ,, 1909.5.4.8-10 1906.5.30.275-279 1959.4.28.281-290 1959.4.28.296-300 1959.4.28.301-303 1959.4.28.307-309 1959.4.28.312-313 1959.4.28.314 1959.4.28.316 1909. ii . 15.40 J 959 .4.28. 274-280 1959.4.28.356 1959.4.28.291-295 1959.4.28.310-311 Locality Uganda Bunjako Jinja Sesse Isls. Bunjako Pilkington Bay Ekunu Bay Pilkington Bay Buka Bay Entebbe, Harbour Macdonald Bay o 4 ' S., 33 14' E. Kenya Kisumu South of Port Southby Kavirondo Gulf Tanganyika Majita Beach near Majita Collector Degen Bayon Degen E.A.F.R.O. A. B. Percival E.A.F.R.O. M. Graham E.A.F.R.O. 28o A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES SUMMARY 1. Eleven species are redescribed on the basis of new material. 2. In addition, three new species (Haplochromis theliodon ; H. empodisma and H. saxicold) are described. 3. Several phyletic lines are represented amongst these fourteen species, which include H. cinereus, a species once thought to represent one of the basic types from which the present-day species-flock had evolved. Evidence now available suggests that, anatomically, H. cinereus is not sufficiently generalized to retain this distinction. 4. Notes are given on the feeding habits and bionomics of the species. 5. Three species (H. obtusidens, H. pharyngomylus and H. ishmaeli) are largely or entirely mollusc-eaters ; three others (H. humilior, H. theliodon and H. riponianus) feed on both insects and molluscs ; one, H . martini, is a piscivorous predator and seven others are insectivore/omnivores. 6. One species, H. niloticus (nom. nov. for Tilapia bayoni Blgr. 1911) is known only from the Victoria Nile, whilst H. ishmaeli occurs in both Lakes Victoria and Edward ; H. humilior is found in Lake Victoria and the Victoria Nile. 7. The assumed distribution of H. macrops in both Lakes Victoria and Edward is discussed ; no definite conclusion can be drawn from the material now available. 8. Two groups of specimens are dealt with in separate appendices because of their uncertain taxonomic position. One group is apparently allied to or even conspecific with H. pharyngomylus and the other with H. ishmaeli. No conclusion can be reached on the status of these aberrant fishes. ACKNOWLEDGMENTS It is with great pleasure that I acknowledge my thanks to Dr. Ethelwynn Trewavas for her advice and criticism and for reading the manuscript of this paper. My thanks are also due to Mr. C. C. Cridland of E.A.F.R.O. Jinja, Uganda, for identifying many of the Mollusca recovered from gut contents. REFERENCES BROOKS, J. L. 1950. Speciation in ancient lakes. Quart. Rev. Biol. 25 : 131. GRAHAM, M. 1929. A report on the Fishing Survey of Lake Victoria, 19271928, and Appendices. Crown Agents, London. GREENWOOD, P. H. 1951. Evolution of the African Cichlid fishes ; the Haplochromis species flock in Lake Victoria. Nature, London, 167 : 19. - 1954. On two cichlid fishes from the Malagarazi River (Tanganyika), etc., Ann. Mag. nat. Hist. (12), 7 : 401. - I956a. The monotypic genera of cichlid fishes in Lake Victoria. Bull. Br. Mus. nat. Hist., Zool. 3 : No. 7. 19566. A revision of the Lake Victoria Haplochromis species (Pisces, Cichlidae). Part I. Ibid, 4 : No. 5. 1957- A revision of the Lake Victoria Haplochromis species, etc. Part II. Ibid. 5 : No. 4. 1959. The monotypic genera of cichlid fishes in Lake Victoria, Part II, and A revision of the Lake Victoria Haplochromis species, etc. Part HI. Ibid. 5 : No. 7. A REVISION OF THE LAKE VICTORIA HAPLOCHROMIS SPECIES 281 PAPPENHEIM, P. & BOULENGER, G. A. 1914. Fische. Wissenschaftliche Ergebnisse der Deutschen Zentral-Afrika Expedition, 19071908, 5 : 225. POLL, M. 1939. Poissons. Exploration du Pare National Albert, mission G. F. de Witte (1933-1935) > f asc. 2 : 78. - & DAMAS, H. 1939. Poissons. Exploration du Pare National Albert, mission H. Damas . fasc - 6 : 73- REGAN, C. T. 1921. The cichlid fishes of Lakes Albert Edward and Kivu. Ann. Mag. nat. Hist. (9), 8 : 632. TREWAVAS, E. 1933. Scientific results of the Cambridge expedition to the East African lakes, 1930-1931. II. The cichlid fishes. /. Linn. Soc. (Zool.), 38 : 309. - 1938. Lake Albert fishes of the genus Haplochromis. Ann. Mag. nat. Hist, (n), 1 : 435. ON A STYGIOMYSIS FROM THE WEST INDIES, WITH A NOTE ON SPELAEOGRIPHUS (CRUSTACEA, PERACARIDA) ISABELLA GORDON BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 5 LONDON: 1960 BY ISABELLA GORDON, D.Sc, Ph.D. Pp> 283-324. With i Table, Plates 3-4 and 34 Text-figures. BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 5 LONDON : 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 5 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued March, 1960 Price Fifteen Shillings ON A STYGIOMYSIS FROM THE WEST INDIES, WITH A NOTE ON SPELAEOGRIPHUS (CRUSTACEA, PERACARIDA) By ISABELLA GORDON, D.Sc., PH.D. SYNOPSIS Stygiomysis holthnisi (Gordon), the second known species of the aberrant Mysidacean genus Stygiomysis, is described and figured in detail. The systematic position of the family Stygio- mysidae is discussed. It is most closely allied to the family Lepidopidae and both families possess certain characters of the suborder Mysida, others of the suborder Lophogastrida. The para- gnatha of some other Mysidacea are figured for comparison with those of Stygiomysis. The termin- ology of the thoracic limbs, and of the segments of the thoracic endopodite, is briefly discussed. A note on Spelaeogriphus lepidops Gordon is added, together with a few comments on the systematic position of the order Spelaeogriphacea. CONTENTS Page I. ON Stygiomysis holthuisi (GORDON) FROM THE WEST INDIES : Introduction . . . . . . . . . .285 Diagnosis of the genus Stygiomysis . . . . . .286 Description of Stygiomysis holthuisi (Gordon) . . . . .287 II. ON THE SYSTEMATIC POSITION OF THE STYGIOMYSIDAE . . . 299 Diagnosis of the family Lepidopidae ...... 308 III. NOTE ON THE PARAGNATHA IN SOME MYSIDACEA .... 309 IV. THE TERMINOLOGY OF : (a) the thoracic appendages . . . . . . . .315 (b) the segments of the thoracic endopodite ..... 319 V. ADDITIONAL NOTE ON Spelaeogriphus lepidops GORDON . . . 320 I. ON STYGIOMYSIS HOLTHUISI (GORDON) FROM THE WEST INDIES. INTRODUCTION. DR. L. B. HOLTHUIS of the Rijksmuseum van Natuurlijke Historic, Leiden, kindly sent me for study three specimens of an interesting cavernicolous " shrimp " which he collected in February, 1957 on the island of St. Martin, Lesser Antilles. When I first examined these specimens I thought that, in spite of the long vermiform body and very short carapace, they looked remarkably like Mysidacea. Then, when I discovered some of their most striking peculiarities, I decided that they did not belong to the Order Mysidacea as at present defined. I fully intended to look up all the literature relating to cavernicolous Mysidacea and, had I done so, I certainly would not have referred these specimens to a new genus. But, when I wrote the preliminary note to N attire (Gordon, 1958), I did not realise that I had inadvertently ZOOL 6, 5 2 o 286 ON A STYGIOMYSIS FROM THE WEST INDIES omitted to refer to Caroli's 1937 paper on the genus Stygiomysis. I have to thank Dr. H. Nouvel of Toulouse for calling my attention to this lapse and I fully agree with him that my genus Rhopalonurus is a synonym of Stygiomysis Caroli. The discovery of a second species of Stygiomysis so far from the locality of the type species of the genus, which is in Italy, is very interesting. Perhaps this rare genus is, or was at some remote period, widely distributed. Caroli had only four specimens of Stygiomysis hydruntina when he published his " nota preliminaire " in 1937. As far as I know, he has never published a fuller description of the species, nor has it been collected again although the two Italian grottos must in all proba- bility have been visited since 1937. I made a fairly complete study of the three type specimens of Stygiomysis holthuisi and had finished the description and figures before I read Caroli's Stygiomysis paper. By that time I also had come to the con- clusion that the species was in fact an aberrant member of the Mysidacea, thus confirming Caroli's opinion. So far no ovigerous female of either species has been found, although in each case a somewhat immature female with only four pairs of developing oostegites is known. Caroli thought it most probable that in Stygiomysis the full complement of oostegites will be seven pairs ; within the order Mysidacea there are as a rule either seven pairs of oostegites, or else two or three pairs Thalas- somysis established in 1939 has four pairs, the first two of which are small, the last two pairs forming most of the brood pouch. But, as Caroli rightly points out, in no Mysid are oostegites absent from the last pair of thoracic limbs. If Stygiomysis should prove to have only four pairs, then they would be absent from the last pair of thoracic appendages and quite exceptional. Recently Dr. Holthuis sent me another specimen from the type locality ; this proved to be a male and the second pair of pleopods differs appreciably from that of the female. Caroli says tkat in Stygiomysis hydruntina none of the pleopods " mostra particolari modificazione sessuali ". This admittedly rather slight sexual dimorphism of the second pair of pleopods is similar to that found in all three monotypic genera of the cavernicolous family Lepidopidae (previously Lepidoph- thalmidae), to which the family Stygiomysidae appears to be most nearly related. After the manuscript was finished, I received from the Amsterdam Museum a fifth specimen which Dr. J. H. Stock obtained in the same place on St. Martin on 4th February, 1959. Order MYSIDACEA Genus STYGIOMYSIS Caroli 1937 Stygiomysis Caroli, 1937, P- 22 - Rhopalonurus Gordon, 1958, p. 1552. DIAGNOSIS. Aberrant cavernicolous Mysidacea with body rather vermiform, somewhat depressed. Carapace short, smooth, fused with tergites of first four thoracic somites, not produced antero-laterally beyond the minute rostral apex and lacking the usual postero-lateral " wings " so that thoracic somites 5 to 8 are entirely exposed. Division of the long thorax into subtagmata further emphasized ON A STYGIOMYSIS FROM THE WEST INDIES 287 by the modification of the first four pairs of thoracic limbs to assist in feeding (maxil- liped and gnathopods see p. 315), the last four pairs being slender peraeopods. Abdomen widest across third somite, sixth somite shorter than fifth. Telson almost as wide as long ; spines on truncate posterior margin only. Ocular scales broader than long, without trace of pigment or visual elements. Antennula more robust and rather longer than antenna, the inner flagellum slightly longer than carapace. Antennal scale minute ; both segments of antennal protopodite distinct (peduncle therefore four-segmented). Labrum broadly triangular. Labium with long, pendant, widely separated paragnatha which conceal part of mandibles. Mandibles with well developed incisior, poorly developed molar, process and a distinct spine-row ; a number of long simple spines on minute third segment of the palp. Maxillula without reflexed palp, distal endite unusually large and robust, the anterior apical spine very large. Maxilla beset with strong spines rather than setae ; the two segments of endopodite very short. Maxilliped (first thoracic limb) pediform, without endites or lobes ; exopodite a small unsegmented lamella ; epipodite rather short and broad. Thoracic appendages 2 to 8 each with well developed multiarticulate exopodite ; endopodites 2 to 4 modified as gnathopods, forwardly directed and stronger than those of the peraeopods (5 to 8). Penultimate segment of thoracic endopodites not subdivided. Pleopods all reduced, biramous, with one-segmented endopodite and three-segmented exopodite (exopodite two-segmented in male pleopod 2) ; alike in both sexes, apart from the sexual dimorphism exhibited by pleopods 2 in Stygiomysis holthuisi. A transverse chitinous lamella, or pair of lamellae, with which the protopodites of the pleopods are incorporated, projects from the posterior sternal margin of abdominal somites 3 to 5. Uropods unusual in having a long, stout, backward prolongation of the protopodite, armed with strong curved spines ; statocyst absent. Oostegites at least four pairs, but the full complement is unknown (Caroli thinks it may be seven pairs). Type species : Stygiomysis hydruntina Caroli 1937, from Italy. A second species from the West Indies is described below. Stygiomysis holthuisi (Gordon) Rhopalonurus holthuisi Gordon, 1958, p. 1552. MATERIAL. The holotype, a female with developing oostegites measuring 9 mm. in length ; the larger paratype is a slightly immature female without oostegites, the smaller paratype is probably immature. 1 In addition to the three original specimens, I received from the same locality at a later date, a male measuring 10 mm. in body length (also in the Leiden Museum). LOCALITY. Devil's Hole, near Simson's Bay (or Simpson's Bay), in the Dutch part of St. Martin, Lesser Antilles, loth February, 1957, (the three types) and, 25th May, 1958 (one male). 1 The holotype and smaller paratype are in the Leiden Museum Collection, and are registered as Crust. G. 434 and Crust. G. 435 respectively. The larger paratype is in the British Museum (reg. no. I959.vii.3.i). 288 ON A STYGIOMYSIS FROM THE WEST INDIES HABITAT. " Devil's Hole lies to the south east of Simson's Bay bridge near the southern shore of St. Martin, just east of Simson's Bay. It is a sink-hole about 5 m. deep in a limestone slope. The hole itself is wide and rather easily accessible. From the hole a few short dead-ending passages run inwards. In one of these passages, which is about 10 m. long, there is a small puddle, about 30 cm. in diameter and 10 to 20 cm. deep. A film of limestone particles covered the larger part of the surface, and a limestone sediment on the bottom made the water turbid when stirred. The water tasted saltish, its salinity being 8100 mg. Cl/litre. It was not in total darkness as the entrance of the passage could still be seen. The distance of the sink-hole from the sea is about 300 m." Dr. Holthuis returned to the same pool two days later, but did not obtain any specimens of Stygiomysis holthuisi, although he got a few Amphipods which he had not seen during his first visit to the cave. Dr. P. Wagenaar Hummelinck, who collected in the same locality on 26th July, 1955, obtained only Amphipods and Nematodes. (Information supplied by Dr. L. B. Holthuis.) ADDITIONAL MATERIAL. A fifth specimen was received after the manuscript was finished. It was found in a small hole or pitch a few inches square, in the type locality, Devil's Hole, St. Martin, by Dr. J. H. Stock on 4th February, 1959. It is a slightly immature specimen, probably female, without oostegites, and is in the Amsterdam Museum. DESCRIPTION OF FEMALE. The largest specimen, measuring 9 mm. in length, and selected as the holotype, is represented in Plate 3. The long, narrow, rather vermiform body is flattened dorsoventrally and is fully segmented behind the carapace, which is just over one- fourth of the total length. The ratio of length to maximum width of carapace is 1-6 : i and the surface is smooth (with perhaps the merest hint of the cervical furrow). The carapace is completely fused dorsally with the tergites of thoracic somites i to 4, being free only at the sides to enclose long narrow branchial chambers. The anterior margin, on either side of the median triangular apex, which scarcely merits the name of rostrum is rather oblique and slightly concave. In front of each concavity is a short broad scale, presumably an ocular scale without any trace of pig- ment or visual elements, ((PI. 3 and Text-figs. 1-3, o). This scale curves downwards and inwards as shown in Text-figs. 2 and 3. The posterior margin of the carapace does not overlap any part of the fifth (first free) thoracic somite ; indeed it is slightly overlapped by the latter in the mid-dorsal line (Text-fig, i and PL 3, 5). The four free thoracic somites pass imperceptibly into the abdomen, although the two tagmata can easily be recognized by their respective appendages. Together they form a very elongated barrel (in dorsal aspect) , widest across abdominal somites 2 and 3 ; abdominal somites 4 and 5 are longest and equal. The telson is as long as the fifth and sixth abdominal somites together, and is nearly as broad as long. The lateral margins are only slightly convergent and the wide posterior margin is armed with strong spines namely, three at each side and a median one flanked by a pair of shorter spines ; four or five spinules fill the gaps between lateral and median spine groups (PI. 3 ; only the median group is visible in Text-fig. 8). ON A STYGIOMYSIS FROM THE WEST INDIES 289 The antennula is longer, and much stouter, than the antenna and is biramous (Text-fig, i and PI. 3). Of the three segments in the protopodite the second is the longest, the first or basal one the widest. The flagella are only slightly unequal in length, the inner one is the more slender, with about 28 segments whereas the outer FIGS. 1-3. Stygiomysis holthuisi (Gordon). Fig. i. Larger paratype, anterior half of body in lateral aspect, after removal of the mouthparts and some thoracic appendages. Fig. 2. Antero-lateral angle of carapace, basal segment of antennula and antennal peduncle, part of labrum. Fig. 3. Right anterior portion of body, in ventral aspect. /. Paragnath of lower lip. m. Mandible, o. Ocular scale, s. Antennal scale, u. Upper lip or labrum. 5. Fifth (first free) thoracic somite, i. First abdominal somite. ZOOL 6, 5 2 o 2QO ON A STYGIOMYSIS FROM THE WEST INDIES one has only 16 to 18 longer segments most of them with long sensory hairs or aesthetascs on the inner margin. The antennulae arise beneath the ocular scales and are not widely separated. The antenna is shorter and much more slender than the antennula. The peduncle of the right antenna of the larger paratype is represented, highly magnified, in Text-fig. 2. The protopodite seems to have only two segments ; the basal one is nearly twice as long as wide, the second one is short and bears, at its outer distal angle, a minute exopodite or scale (s). The two basal segments of the endopodite are each long and slender (at least four times as long as wide) and the flagellum, which has 16 to 18 segments, is about four-fifths of the carapace length (PL 3). Mouthparts. Those on the right side of the larger paratype were removed for detailed examination. Each appendage was sketched from various aspects before any attempt was made to stain and mount it. In this way any distortion due to movement of, or pressure from, the coverslip was eliminated. Certain details not apparent in the unstained appendages were inserted later from the micro-prepara- tions. The labrum or upper lip is large and triangular, with rounded apex, beyond which, at a lower level, a small rounded lobule is visible (i.e. the lobule is more dorsal in position) see Text-figs, i and 3, u. Behind the mandible is the labium or lower lip which is unusual in that the paragnatha are widely separated ; moreover, each is enormous, pendant and forwardly directed so as to overlap and partially conceal the mandible (Text-figs, i and 3, /). The right mandible is shown in situ in Text-fig. 3 and the distal portion is represented in Text-fig. 4. The crown consists of a well developed, three-lobed incisor process (with a subsidiary tooth at the base of one of the lateral ones) a small but well chitinized knob-like molar process and, between the two processes, a strongly chitinized triangular lobef and a row of five graded spines. The palp has the usual three segments ; the median one is very long and somewhat twisted, the short terminal one has seven long spines. The maxillula is unusually large and massive (Text-fig. 5). The proximal endite, Hansen's lobe i, has four terminal spines, two long and plumose, two short and simple. The distal or major endite, Hansen's lobe 3, has a terminal row of spines the anterior one unusually long and stout and, near its base, is a long seta. There is no backwardly directed endopodite or palp. The maxilla is a thin leaf-like appendage abundantly beset with spines and setae (Text-fig. 6) . Hansen's pi. 3 was not apparent until the appendage had been stained and mounted. The proximal endite (s) , in this case the lobe from 2, is simple, with a median seta on the posterior margin and a series of apical and subapical spines. The endite from pi. 3 (d) is divided into a small distal lobe with four terminal spines, and a large oval proximal lobe with a row of nine marginal spines. A short, two-segmented endopodite (en) follows ; the distal segment bears three long spines, the proximal one has two spines on the inner margin. A long narrow exopodite (ex] arises from the outer margin of pi. 3 ; its margins, the free inner one excepted, bear plumose setae. The first thoracic appendage or maxilliped is represented in Text-fig. 7 at a lower magnification than the other mouth parts. Though much smaller than, it is essenti- t The fixed lacinia, against which the lacinia mobilis works. A STYGIOMYSIS FROM THE WEST INDIES 291 ally similar to, the three following pairs of thoracic limbs. However, it alone bears a fairly large epipodite reflexed into the brancial chamber and, moreover, the exopo- dite is reduced to a small oval unsegmented lamella (which may also be respiratory). FIGS. 4-6. Stygiomysis holthuisi (Gordon). Mouthparts of larger paratype. Fig. 4. Distal portion of right mandible, with palp. Fig. 5. Right maxillula. Fig. 6. Right maxilla. All at higher scale except upper Fig. 4. d. Divided endite. en. Endopodite. ex. Exopodite. 5. Simple or undivided endite. 2 and 3 Hansen's plates 2 and 3 of protopodite (coxa and basis) . 2Q2 ON A STYGIOMYSIS FROM THE WEST INDIES The endopodite a continuation of the axis of the protopodite comprises five segments and a strong terminal claw ; the " knee " joint between the third and fourth segments is indicated by an arrow. There is a row of strong spines on the inner margin of segments 2 and 3, a group of spines on the distal half of the inner surface of segment 4 and a pair of claw-like spines, flanking the terminal claw, on segment 5. (Since the terminology of the segments of the endopodite varies, I have thought it better in the description of the thoracic limbs to use numbers see p. 319 for discus- sion of this). 8 FIGS. 7 and 8. Stygiomysis holthuisi (Gordon). Fig. 7. Right first thoracic limb (maxilliped) of larger paratype. Fig. 8. Posterior part of body of holotype, in ventral aspect, e. Exopodite. b. Basis, ep, Epipodite. m. Membranous free lamina connecting and incorporating protopodites of pleopods 5. t. Telson. p. Backward prolongation of protopodite of uropod. 5 and 6. Fifth and sixth abdominal somites. 3 and 4. Segments of endopodite on either side of " knee " joint (indicated by arrow). Thoracic appendages 2 to 8. Each of these bears a well developed, natatory exopo- dite, as in most Mysidacea. All are visible in dorsal aspect, as represented in PI. 3, each is composed of a long robust proximal segment followed by a more slender tapering flagellum of 10 to 12 segments (Text-figs. 9, 10, n, e). Each segment of the flagellum carries a pair of plumose setae, but only the base of each is inserted in Text-figs. 9-11. These seven pairs of appendages are arranged in two distinct groups. The three anterior pairs have their endopodites directed forwards, more robust than, and obvi- ON A STYGIOMYSIS FROM THE WEST INDIES 293 FIGS. 9-1 1. Stygiomysis holthuisi (Gordon). Larger paratype. Fig. 9. Right second thoracic limb (first gnathopod) . Fig. 10. Fourth thoracic limb. Fig. u. Eighth thoracic limb. b. Basis, e. Exopodite. i to 6. Segments of endopodite. " Knee " joint indicated by arrow. 294 ON A STYGIOMYSIS FROM THE WEST INDIES ously differing in function from, those of the four posterior pairs which are slender ambulatories. The three anterior pairs should be termed gnathopods since they are probably used to seize prey and pass it to the mouthparts. Thoracic appendages 2, 4 and 8 were removed from the right side of the larger paratype and are repre- sented in Text-figs. 9, 10, and n respectively. Appendage 2 is, however, represented in inner, the other two in outer, aspect. The endopodite, beyond the basis which also bears the exopodite (e), comprises, as already mentioned in connection with the maxilliped, five segments, or six if the terminal claw is included (Text-fig, n, J-6). The endopodite of gnathopod i (thoracic appendage 2) is very robust ; segments i and 2 are short, 3 is the longest and by far the broadest, 4 is long and armed with a longitudinal series of six setae on the inner surface and 3 on the ventral margin (with another near the margin), segment 5 is not quite half as long as 4 and is armed with a number of short spines in addition to the terminal claw. Gnathopod 2 (appendage 3) is very similar to, but a trifle less robust than, gnathopod i and there are only 4 long setae on the distal half of the inner surface of segment 4. As illustrated in Text-fig. 10, the third gnathopod is much more slender than gnathopod i, with only three setae on the distal third of the inner surface of segment 4. Thoracic appendages 4 to 8 are all very similar ; they are in fact peraeopods or walking legs, with the endopodite slender as represented in Text-fig, n. Segment 4 is the longest but segment 5 and the terminal claw are also elongated so that the portion distal to the " knee " is longer than the proximal part. On the ventral margin of segment 5 is a row of curved serrated spines. The pleopods are similar in form and do not differ much in size from the first to the fifth pairs. Each is a small, biramous appendage comprising a short protopodite, a one-segmented endopodite and a three-segmented exopodite (Text-figs. 8, 14 and 1 6). The first pair are more widely separated than are any of the others, i.e. they are more laterally placed, so that in dorsal aspect even the endopodites are visible (PI. 3 and Text-fig. 14) ; in ventral aspect they are concealed by the last pair of thoracic appendages. In other words, abdominal somites 2 to 5 are considerably deeper than somite i and their tergites bend in somewhat to meet their respective sternites, thus the pleopods are more ventral in position. A most unusual feature is the presence of a transverse lamella projecting freely backwards from the posterior sternal margin of each abdominal somite 3, 4 and 5. This chitinous duplicature, as Caroli terms it, becomes progressively longer (that on somite 3 being the shortest, that on 5 the longest). The protopodites of each pair of pleopods are incorporated with the respective lamella, but the muscle bands which work the two rami of each pleopod are visible through the chitin (Text-figs. 8, 14 and 15 m). Under certain lighting conditions, at a certain angle, there appears to be a median notch (an inverted V) dividing the lamella on somite 3 of the smaller paratype ; when a fine needle point is run along under the free margin of the lamella, however, the posterior margin is found to be continuous. One of the most distinctive features of this animal is the uropod (Text-figs. 8, 12 and 13). The protopodite is expanded medially (e), beyond the base of the endo- podite (en), so that it occupies very nearly half of the posterior margin of the sixth abdominal somite (Text-fig. 8). Moreover, in ventral aspect, it has a long, massive, ON A STYGIOMYSIS FROM THE WEST INDIES 295 13 14 FIGS. 12-14. Stygiomysis holthuisi (Gordon). Smaller paratype. Fig. 12. Left uropod, in dorsal aspect. Fig. 13. Same, in ventral aspect. The muscle bands within the proto- podite are restricted to the stippled area. Fig. 14. Abdomen and telson, in lateral aspect. e. Median expansion of protopodite. en. Endopodite. ex. Exopodite. m. Membranous lamellae connecting protopodites of pleopods 3, 4 and 5. p. Backward prolongation of protopodite. i and 6. First and sixth abdominal somites. 296 ON A STYGIOMYSIS FROM THE WEST INDIES backward prolongation (p), which extends almost to the posterior margin of the endopodite and bears five strong apical and subapical spines. These spines curve inwards and upwards (dorsally) towards those on the other partner. On the distal half of the inner margin of p there is a row of spinules interrupted by one or two spines. The left uropod of the smaller paratype was removed and is represented, in dorsal and ventral aspects, in Text-figs. 12 and 13. In dorsal aspect only the proxi- 15 FIGS. 15-16. Stygiomvsis holthmsi (Gordon). Fig. 15. Third left pleopod of female, and left half of membranous lamella, in ventral aspect. Fig. 16. Second pleopod of holotype (female), m. Lamella. 3. Side of third abdominal somite. Fig. 16 at larger scale. mal part of the protopodite with its medial expansion (e) is visible. The endopodite conceals part of the exopodite, and the long marginal plumose setae on exopodite and endopodite (mostly omitted in Text-fig. 12) conceal the apical spines on the prolongation p. There is no suture line or diaerisis on the exopodite, at the level of the spines on the outer margin, marking off the distal portion. The uropods are unusually stiff (at any rate in fixed specimens) ; when I tried to spread out the exopodite ON A STYGIOMYSIS FROM THE WEST INDIES 297 from beneath the endopodite it was difficult to do so without damage to the exopodite base (Text-fig. 13 ex). There is no statocyst either in the uropod or in the basal segment of the antennular peduncle. Oostegites. In the holotype there are four pairs of developing oostegites arising from the coxae of thoracic limbs 4 to 7 respectively. The two anterior pairs are equally developed, each partner in the form of a long narrow lamella. The third pair is unequal, the left partner being appreciably larger than the right one ; of the last pair the left one is very small and the right one has not as yet appeared. NOTE ON THE MALE. On 27th October, 1958 I received from Dr. Holthuis another specimen which had been collected in the type locality on 25th May, 1958 by Mr. M. Vroman. This specimen, which proved to be a male, is rather more slender than the holotype and measures 10 mm. in body length. It is less contracted than the type specimens and not quite so flattened dorsoventrally (Text-fig. 24). The ratio of carapace length to total body length is perhaps more as in the living animal (i : 474 in the male, as against i : 3-93 in the holotype). Apart from the more slender build, the presence of short penes, and the sexual dimorphism exhibited by the second pair of pleopods, I find no obvious differences between the male and the holotype the antennula is often modified in male Mysidacea. Pleopod 2 is repre- sented in Text-fig. 18, that of the holotype in Text-fig. 16. As far as I can see without actually removing one of these pleopods from each specimen, the exopodite is three- segmented in the holotype and only two-segmented in the male. In the latter there are 8 or 9 long, strong, thorny setae (or spines) on the inner margin of the proximal segment and, in addition, 5 to 7 shorter setae which are more oblique (directed more ventrally) see Text-fig. 18. The first pleopod is represented in Text-fig. 17. The genital opening is situated on a very low papilla at the base of the last thoracic appendage, on the inner side of the coxa. There is no coxal lamina such as Villalobos (1951, p. 208, Text-fig. 12) has figured for his genus Typhlolepidomysis. REMARKS. These specimens from the West Indies are undoubtedly referable to Caroli's genus Stygiomysis, as a comparison of Text-fig. 8 with his Text-fig. 3 (1937, p. 223) shows. The type species of the genus, 5. hydruntina Caroli, was found in two grottos in the region of Otranto, Italy (the male holotype in the " Buco dei Diavoli ", one female and two immature specimens in " 1'Abisso "). Caroli's brief description, based chiefly on the holotype, makes known the most important charac- ters of the species, and indeed of the genus. As far as I know, the more detailed study which was to have followed (Caroli, 1937, p. 220) has never been published. In addition to the photograph of the whole animal in lateral aspect, Caroli figured only the first gnathopod (second thoracic appendage) and the posterior end of the body in ventral aspect. Most of his description of S. hydruntina would apply equally well to 5. holthuisi. Nevertheless, the two species differ in certain respects. For example, in S. holthuisi segment 5 of the endopodite of gnathopod i (that which bears the terminal claw) is much longer than in 5. hydruntina and the terminal claw and the pair of spines flanking it are very much shorter (in hydruntina they are much longer than the segment that bears them cf. Text-fig. 9 with Caroli, 1937, fig. 2, p.. 222) . In S. holthuisi the terminal claw of peraeopod 4 (last thoracic limb) is rather shorter than segment 5, whereas in 5. hydruntina the claw is longer than the dactylopodite. ZOOL. 6, 5 2o 298 ON A STYGIOMYSIS FROM THE WEST INDIES The shape of the uropod differs appreciably in the two species ; in S. hydruntina the outer margin of the prolongation p of the protopodite is distinctly concave, the exopodite is much broader, with a more semicircular apex (cf. Text-figs. 8, 12 and 13 with Caroli, 1937, p. 223, fig. 3). According to Caroli, the pleopods of the male increase very slightly in length from FIGS. 17-18. Stygiomysis holthuisi (Gordon). Male. Fig. 17. First pleopod. Fig. 18. Second (modified) pleopod. the first to the fifth pair, but " nessuno di essi mostra particolari modificazioni sessuali ". In S. holthuisi, however, the second pair of pleopods in the male are modi- fied ; as far as I can ascertain without removing one of these pleopods, the exopodite has only two, instead of the usual three, segments, and the proximal segment has a row of 9 long spines as represented in Text-fig. 18 (compare with the first male pleopod, Text-fig. 17 and with the second female pleopod, Text-fig. 16). Also, the membranous flaps between the third, fourth and fifth pairs of pleopods differ in ON A STYGIOMYSIS FROM THE WEST INDIES 299 the two species. Carol! says that, in each of these three pairs, " the peduncle (proto- podite) is completely fused with a wide duplicature of the integument, in the form of an almost semicircular lamina, movably articulated with the posterior margin of the sternite, [extending] from the external margin to the middle line, where it meets that of the other side " (my translation). He adds that the peduncle proper is indicated by the presence of the muscle bands which work the two rami, but that there is no suture between it and the duplicature (Caroli, 1937, pp. 223-224 and fig. 3). In S. holthuisi these duplicatures unite to form a single membranous flap, or lamina in which the muscle bands are visible at either side (Text-figs. 8 and 15). I thought at first that there was a median slit separating the two halves of the mem- brane of the third pair of pleopods of the smaller paratype of 5. holthuisi. But, on re-examination, I find that there also the membrane is continuous although it is slightly creased in the middle (Text-fig. 14). The female specimen of 5. hydmntina measured only 7 mm. in length, but it also had 4 pairs of developing oostegites arising from thoracic limbs 4-7. These oostegites were probably not quite so well developed as those in the holotype of S. holthuisi (c.l. 9 mm.). Caroli is almost certain that the full complement of 7 pairs will be found in Stygiomysis. If not, then Stygiomysis will be quite exceptional. Seven pairs of oostegites are found in the Lophogastrida (Lophogastridae and Euco- piidae), in the Lepidopidae (all three genera), Petalophthalmidae and Boreomysinae. All the rest of the Mysida (Thalassomysis excepted) have only two or three pairs of oostegites arising from the last two or three pairs of thoracic appendages. It is to be hoped that an ovigerous female of either species of Stygiomysis will soon be found, so that the exact number of oostegites can be ascertained. II. ON THE SYSTEMATIC POSITION OF THE ST YGIOM YSI D AE Up to the present time the following cavernicolous Mysidacea have been described or mentioned in the literature : STATOCYST ABSENT Lepidopidae : Lepidops servalus (Fage) 1924, 1925. Zanzibar. Spelaeomysis bottazzii Caroli 1924. Italy. (Found a second time by Stammer see Caroli 1937, p. 226). Typhlolepidomysis quinterensis Villalobos 1951. Mexico. Stygiomysidae : Stygiomysis hydruntina Caroli, 1937. Italy. Stygiomysis holthuisi (Gordon) 1958. West Indies. STATOCYST PRESENT Mysidae : Heteromysis cotti Caiman 1932. Lanzarote. Troglomysis vjetrenicensis Stammer 1933, 1936. Herzegovina. Antromysis cenotensis Creaser 1936. Yucatan. Antromysis sp. Bolivar y Pieltain 1943-44, and 1950. Cuba. 300 ON A STYGIOMYSIS FROM THE WEST INDIES Most of these species occur in fresh or brackish water ; Heteromysis cotti is a marine species since the cave in which it lives communicates with the sea. (Antro- mysis anophelinae W. M. Tattersall, 1951, p. 230, was discovered in the holes made by a land crab, Cardisoma crassum Smith, along with mosquito larvae, far up the mangrove inlets behind Puntarenas, Costa Rica in water which must have been nearly, if not quite fresh) . The cavernicolous species with a statocyst in the endopo- dite of the uropod are referred to the large family Mysidae, in the suborder Mysida. Those that have no statocyst are peculiar or aberrant in certain respects and their systematic position is less easy to determine. Fage referred the family Lepidopidae to the suborder Mysida while admitting that it had a number of Lophogastrida characters ; Caroli thought the family might be regarded as modified or aberrant Lophogastrida and Villalobos thought that it should be placed in a new suborder (see p. 306). Caroli had no hestitation in referring the Stygiomysidae to the suborder Mysida. GENERAL APPEARANCE. In all three genera referred to the family Lepidopidae (formerly Lepidophthalmidae) the carapace agrees with the accepted definitions of the order Mysidacea. Caiman (1909, p. 171), in his definition of the order writes " the carapace extends over the greater part of the thoracic region, but does not coalesce with more than three thoracic somites ". Under morphology lower down on the same page he adds " from five to seven of the thoracic somites are distinct, and the last two or three may be left uncovered by the carapace on the dorsal side ". Caiman's definition is adopted by Tattersall, (1951, p. n) and that given by Zimmer (1927, p. 609) is in agreement with it. It has been altered by Dr. Olive Tattersall to read " carapace shield-like, covering the greater part of the thoracic region, fused dorsally with the head region and not more than four anterior somites including the first, which is incorporated in the head " (Tattersall and Tattersall, 1951, p. n). Dr. Tattersall, in a letter dated I2th November, 1958, informs me that, as far as she can make out, " in some Mysidae there does appear to be fusion of at any rate part of the fourth somite with the carapace (in e.g. Mysidopsis, some Siriella and in Neomysis). I do not know quite how much one can rely on the " saddle " on the carapace in so many mysids, but I always considered it to represent the area of fusion of the two surfaces, i.e. the inner surface of the carapace with the dorsal surface of the thoracic somites involved." I have examined specimens of Lepidops servatus (Fage) but they are delicate and none too well preserved, and it is not easy to discover exactly how many thoracic tergites are free. As in most Mysidacea, the anterior thoracic somites are short and rather crowded together, especially dorsally, and the suture lines are visible laterally. As far as I can see four are certainly free and perhaps five. The last two tergites, which are uncovered by the emarginated carapace, are each as long as the first abdominal somite. The large lateral " wings " of the carapace extend to the first abdominal somite, and can easily be moved away from the body. All three genera retain the " caridoid facies " as do most Mysidacea. Stygiomysis, on the other hand, does not agree with Caiman's definition, even as emended by Dr. Tattersall. The carapace is unmistakably fused with the first four thoracic tergites but has no " wings " to speak of, so that the last four thoracic ON A STYGIOMYSIS FROM THE WEST INDIES 301 somites are entirely exposed (PL 3, Text-figs, i and 24). Moreover, the body behind the short carapace (the sides of which cannot easily be separated from the body at any rate in preserved specimens) is vermiform and very flexible. Caroli's photograph of Stygiomysis hydruntina shows the abdomen curved gracefully upwards, in a most un-mysid-like position (1937, p. 221, fig. i). As that author says, the short carapace and the apparent division of the thorax into two subtagmata gives the genus "un aspetto particolare ". The sixth abdominal somite is distinctly shorter than either the fifth or the fourth somite. In Mysidacea " the sixth abdominal somite is usually longer than any of the others " (Tattersall and Tattersall, 1951, p. 16). This is the chief reason why I at first thought that the genus could not be referred to the order Mysidacea as at present defined (Gordon, 1958). The other two most striking peculiarities of the genus Stygiomysis namely, the transverse membranes with which are incorporated the protopodites of pleopods 3, 4 and 5, and the long backward extension of the protopodite of the uropod, are not quite as unique as I had supposed. Each will be dealt with in turn. TRANSVERSE CHITINOUS LAMELLAE. Caroli (1937, p. 225) found what he called homologous though not identical structures in Lepidops and Spelaeomysis. I have been able to confirm his observation in Lepidops ; between the bases of pleopods 3, 4 and 5 there is a transverse laminar extension from the posterior margin of each abdominal somite, as represented in Text-fig. 20. These chitinous duplicatures of the integument, as Caroli calls them, increase in length from somites 3 to 5 but the protopodites of the pleopods are not incorporated with them. In Spelaeomysis they are said to be even better developed, with a median emargination in the posterior border. Caroli assumes from this that each lamina had a double origin i.e. that it arose as two duplicatures of the integument which had fused medially (Caroli, 1937, p. 225, cf. also his fig. 3). Unfortunately, Villalobos (1951) does not state whether or not these laminae are present in Typhlolepidomysis ; they were overlooked when Lepidops and Spelaeomysis were first examined and could easily have been overlooked a third time. (But see footnote to p. 308). Although transverse laminae are known only in these more or less aberrant Mysidacea among the Malacostraca, they recall somewhat analogous structures in the Copepoda. Here it is the thoracic limbs which are involved ; " the proximal segments of the protopodites of each pair are connected with each other across the middle line by a plate formed by a transverse fold of the sternal integument . . . so that in the backward and forward movement in swimming the two appendages move as one " (Caiman, 1909, p. 81). See, for example, the figures of the reduced last limb of male Harpacticids (?5 $) in Chappuis 1931, p. 521, fig. 13 ; p. 538, fig. 60 these limbs are not always joined in the female. How do the duplicatures arise in the Mysidacea? In Lepidops the chitinous lamina seems to be simply a backward extension or fold of the posterior sternal margin of the abdominal somite, since the protopodites of the pleopods are not involved (Text -fig. 20, m). In Stygiomysis hydruntina, Caroli's figure (1937, fig. 3) suggests that each lamina may have a double origin that each half arises as a median expansion of the protopodite of the pleopod which fuses anteriorly with the sternal margin of the abdominal somite, since the two halves merely meet in the 302 ON A STYGIOMYSIS FROM THE WEST INDIES ON A STYGIOMYSIS FROM THE WEST INDIES 303 middle line. In Stygiomysis holthiiisi the lamina might arise in either of these ways, since the protopodites of the pleopods are incorporated and there is no trace of a median suture line (see also p. 294). It would be interesting to know how these chitinous laminae develop and what their function can be, since the pleopods are almost certainly not natatory. The very thin sternal wall above each lamella may be respiratory, UROPOD. In Lepidops servatus the protopodite is broad, occupying half of the posterior margin of the last abdominal somite. Moreover, on the ventral side, there is a short but distinct backward prolongation armed with a terminal spine and with numerous spinules on the inner border (Text-fig. 21, p). This might represent an initial stage in the development of a massive backward prolongation such as is found in Stygiomysis, although Caroli prefers to regard it as a last vestige retained in the family Lepidopidae (1937, p. 225). At any rate, this incipient or vestigial structure in Lepidops (and presumably also present in Spelaeomysis) , together with the transverse laminae mentioned above, suggest a fairly close relationship between the Lepidopidae and the Stygiomysidae (see also p. 308). A somewhat analogous structure is characteristic of the uropod of the Stomatopoda, which might indicate that the prolongation of the protopodite is archaic and, as Caroli thinks, may at one time have been much more widely distributed among Malacostraca. I am inclined to think that it is a highly specialized structure which has arisen indepen- dently in the Stygiomysidae, and to a lesser extent in the Lepidopidae, in response to some special habit. It may be mentioned in passing that, in the genus Micraspides (Syncarida, Anaspi- dacea), the uropod is also armed with strong spines, not on the protopodite (which is not prolonged) but on the endopodite. On the dorsal surface, near the postero- median angle, are three strong spines directed obliquely upwards ; in addition there is a row of special spines on the inner margin (Text-fig. 19, en). Nicholls' figure (1931, pi. 32, fig. 15) of the uropod of the type species Micraspides calmani is rather small and does not show the row of spines near the external margin of the exopodite. In Thermosbaena mirabilis Monod the exopodite of the uropod is remarkably spiny (Monod, 1940, fig. 19). PLEOPODS. The pleopods of Stygiomysis, though smaller and not increasing much in length from the first to the fifth pair, are of the same general form as those of Lepidopidae (PI. 3, and Text-fig. 14, cf . Text-fig. 20) . In both families they are reduced, non-natatory, and biramous in both sexes. In Stygiomysis holthuisi the second pair is modified in the male although it is not longer than the others (Text-figs. 17 and 1 8) ; Caroli's statement that there is no sexual dimorphism of the pleopods in FIGS. 19-21. Micraspides calmani Nicholls. Fig. 19. Left uropod of syntype, in dorsal aspect, omitting long marginal setae. Lepidops servatus (Fage). Syntype from Paris Museum. Fig. 20. Abdominal somites 3 to 6, and protopodites of uropods, in ventral aspect, showing membranous free laminae (m) on posterior margin of somite 3, 4 and 5, between bases of the pleopods. Fig. 21. Left uropod, in ventral aspect (endopodite damaged), en. endopodite. ex. Exopodite. p. Short backward prolongation of proto- podite. Scale = o - 5 mm. 304 ON A STYGIOMYSIS FROM THE WEST INDIES S. hydruntina requires corroboration. The second pleopod of the male is similarly modified in all three genera of the Lepidopidae see Fage, 1925, fig. VI, 2 and 2 <$; Caroli 1924, p. 513 ; Villalobos 1951, fig. 13, A and B. (A glance at fig. i of Villalobos' paper shows that pleopod 2 is shorter than pleopod 5 in the male holotype ; it seems, therefore, that in his fig. 13, C and D are drawn at a smaller scale than A and B). Pseudobranchiae are absent in both Stygiomysidae and Lepidopidae. In Lophogastrida the pleopods of both sexes are well developed, biramous, multi- segmented and natatory ; they may be considerably larger in the male than in the female, but are unmodified or very slightly so. In Mysida they are seldom alike in both sexes and generally exhibit marked sexual dimorphism. In the female they are, as a rule, rudimentary, uniramous (rarely the first may be biramous) and non- natatory ; occasionally those of the male are all rudimentary as in the female (Mysidella, Mysidetes, Heteromysis). In the male they are sometimes biramous, natatory and unmodified (Petalophthalmidae), or the third, or second and third, pairs are slightly modified (Boreomysinae) ; more often one or more pairs are reduced as in the female and one or more pairs are modified as accessory copulatory organs ; pseudobranchiae are present. The pleopods which are most commonly modified are the third and/or fourth pairs. In the genus Rhopalophthalmiis it is the second pair (Tattersall, O.S. 1957, fig. 7H, compare with G, J, K and L) ; but here the pleopods are biramous, multiarticulate, with conspicuous pseudobranchial lobes at the bases of the endopodites. The modified pair, or pairs, of male pleopods are usually long and conspicuous in the family Mysidae, to which Rhopalophthalmus belongs. It is just a coincidence that the same pair of male pleopods is modified in this genus as in the cavernicolous families under discussion. MOUTHPARTS. There are numerous differences in detail between the individual mouthparts of the two families and these are doubtless correlated with differences in feeding habits. For example, in the Lepidopidae the molar process of the mandible is better developed, the palp is larger and the second and third segments are each armed with a close row of thorny spines (cf. Text-fig. 27, where many of the spines and setae are omitted, Fage, 1925, fig. IV, M and Villalobos, 1951, fig. 6 with figs. 4 and 5). According to Tattersall and Tattersall, (1951, pp. 20-21), in Mysidacea the maxil- lula is smaller than the maxilla and always lacks a backwardly directed palp or endopodite except in the genus Gnathophausia. This statement requires amplifica- tion however, because there is a palp in all three genera of the Lepidopidae (Fage, 1925, fig. IV, w 1 ; Caroli, 1924, p. 513 and Villalobos, 1951, fig. 7). In each of three specimens of Lepidops servatus which I have examined, this palp has a median constriction and appears, therefore, to be two-segmented as in Typhlolepidomysis (see Text-figs. 27 and 28). Moreover, in Sfygiomysis holthuisi the maxillula is larger than the maxilla ; the outer endite is stout and one of its apical spines is greatly enlarged ; there is no palp (Text-figs. 2 and 5 ; cf . the small maxillula of Lepidops represented in Text-figs. 28 and 27). The maxilla of Stygiomysis has strong spines rather than setae on the whole inner margin but they do not continue along the lower border of the simple endite (Text- fig. 6, on en, d and s) ; in the Lepidopidae the inner and lower margins of the maxilla ON A STYGIOMYSIS FROM THE WEST INDIES 305 bear numerous finer setae and spines, and the segments of the endopodite are much longer (Fage, 1925, fig. IV, m 2 and Villalobos, 1951, fig. 8). The first thoracic append- age or maxilliped of Stygiomysis is more pediform, lacks the large endites, but has strong spines on the inner margin of the distal segments of the endopodite (Text-fig. 7 ; cf. Fage, 1925, fig. VI, mp and Villalobos, 1951, fig. 9). According to Siewing (1958, pp. 250-251) this appendage should be called a gnathopod in the Stygio- mysidae and a maxilliped in the family Lepidopidae (see later, under terminology, p. 316). Differences such as these may be found between closely related families of Mysidacea. The exopodite of the first thoracic limb is reduced to a simple unseg- mented lamina, which is relatively smaller in Stygiomysis holthuisi than in either Lepidops or Typhlolepidomysis (cf. Text-fig. 7 with Fage, 1925, fig. VI, mp and Villa- lobos, 1951, fig. 9). In this respect both families are nearer to the suborder Lopho- gastrida and to the Petalophthalmidae (in which the exopodite is lacking, as in some species of Gnathophausia) than to the Mysidae. LABIUM. The paragnatha in Stygiomysis are unusual and very conspicuous, although Caroli did not mention them. To discuss the labium at this point would necessitate rather a lengthy digression ; it will be dealt with in a separate section later on (p. 309). Of the paragnatha that I have examined, those of Lepidops seem to be nearest to those of Stygiomysis. THORACIC APPENDAGES 2-4. Many of the peculiarities of the mouthparts men- tioned above are doubtless correlated with the fact that in Stygiomysidae three pairs of thoracic appendages (2-4) are modified as gnathopods, whereas in Lepi- dopidae only one pair, the second, is thus modified (cf. PI. 3, Text-figs. 9 and 10 with Fage 1925, fig. V, p 2 and Villalobos, 1951, fig. 10). Incidentally, Villalobos refers to the second thoracic limb as " primer par de pereiopodos " in the text and as " maxil- lipedo del secundo par " in the legend to fig. 10 and Caroli (1937, p. 226) says " toracopodi 2-4 (gnathopodi) massillipediformi "! Only in the family Eucopiidae of the suborder Lophogastrida are the same three pairs of thoracic limbs modified as gnathopods ; but in this family the following three pairs of appendages are elongated and specially modified (Tattersall and Tattersall, 1951, p. in, figs. 136, D and E and 12 A). When Sars (1885, p. 59) suggested that the long prehensile thoracic limbs 5 to 7 might be used for seizing hold of Hydroids or Crinoids fixed at the sea bottom, he was unaware of the fact that the Eucopias are bathypelagic (Fage, 1941, p. 3 and 1942 pp. 3 and 60). There is nothing in mid-water for a bathy- pelagic animal to cling to ; these prehensile endopodites, reaching well beyond the body, seem admirably suited for keeping the animal suspended and for seizing prey which could be passed to the short robust endopodites of the gnathopods (see p. 319). Stygiomysis is probably raptorial in its feeding habits whereas the Lepidopidae are probably predominantly filter feeders. OOSTEGITES. In Lophogastrida, Lepidopidae, Petalophthalmidae and Boreomy- sinae (Mysidae) a pair of oostegites or brood lamellae arises from each pair of thoracic appendages 2-8. Throughout the rest of the large family Mysidae there are either three or two pairs, arising from the last three or two pairs of thoracic appendages. The only exception to this rule is the genus Thalassomysis Tattersall 1939, p. 238 (established two years after Caroli's genus Stygiomysis). Here reduction has not 306 ON A STYGIOMYSIS FROM THE WEST INDIES gone quite as far and four pairs of oostegites are present. Of these the first pair is vestigial, the second pair small and the last two pairs form the major part of the brood chamber. Oostegites disappear from the more anterior limbs first so that, no matter what their number, a pair is always present on the seventh and the eighth thoracic appendages. In Stygiomysis the full complement is unknown but, unless the genus is quite exceptional, there ought to be a pair on the eighth thoracic limbs ; this means that there ought to be at least five pairs and Caroli may be right in thinking that there may be seven pairs. DISCUSSION. The family Stygiomysidae seems to be distinct from, but most closely related to, the family Lepidopidae. But, as already mentioned, opinion differs as to the systematic position of the family Lepidopidae. Since gills are absent and the pleopods are reduced, non-natatory, Fage (1924, 1925) referred his new family " Lepidophthalmidae " to the sub-order Mysida (Mysidea of some authors), at the same time stressing its affinities with the sub-order Lophogastrida. Both Caroli and Villalobos have raised objections to this. Caroli's announcement of Spelaeomysis appeared before Page's main paper (in December, 1924). He says that the discovery of a second genus in Italy, so far from the habitat of the first genus (Zanzibar), shows that these forms must have had at one time a very wide distribution. " Pero e forse prematuro ritenerle come forme di passagio dai Lopho- gastrida ai Mysida. A mio avviso, potrebbero anche essere considerate Lophogastrida modificati (perdite delle branchie, riduzione dei pleopodi) dalla vita cavernicola assunta in epocha remota " (Caroli, 1924, p. 513). That Fage adheres to his original opinion is evident from a remark he makes in a later work on Lophogastrida (1941, p. 3) ; discussing some primitive characters he mentions " le palpe retro verse, herite des Cumaces et des Tanaidaces, que portent au premieres maxillae les Gnatho- phausia. Seuls parmi les Mysides [Mysida] les genres Lepidops (Lepidophthalmtis Fage 1924) et Spelaeomysis Caroli 1924 en sout pourvus, qui revelent ainsi leur part du meme heritage ". Villalobos (1951, p. 216), after a brief discussion of the syste- matic position of the Lepidopidae concludes : " Segun nuestro punto de vista, la position adoptada por Fage es un poco forzada, porque si los caracteres distinctivos de la familia se encuentran representados ya sea en Lophogastridea o en Mysidea, lo conveniente seria crear un soborden intermedio ". He did not, however, propose a new sub-order because of the paucity of material and the small number of genera and species known (three monotypic genera). When Caroli (1937, p. 225) discovered the transverse lamellae and a short but distinct backward prolongation of the protopodite of the uropod in Lepidops and Spelaeomysis (see Text-fig. 20, m and Text-fig. 21, p of this paper) he said that he could not judge whether the presence of these two characters indicated a closer affinity between the Lepidopidae and Stygiomysis. " Si pud anche suppore che si trata di caratteri arcaica, communi probabilmente in passato ad altre e piii numerose forme, e conservatisti, benche in via di graduale riduzione, nei nostri Misidacei, in grazzia dello speciale habitat ". He does not explicitly say that he has altered his earlier opinion (1924) namely, that the Lepidopidae are perhaps modified Lopho- gastrida, and he has no hesitation in referring the Stygiomysidae to the sub-order Mysida. ON A STYGIOMYSIS FROM THE WEST INDIES 307 The family Lepidopidae shares its diagnostic characters with the rest of the Mysi- dacea as follows : (i) Gills absent : Stygiomysidae and the whole of the Mysida. (ii) Oostegites seven pairs : Lophogastrida, Petalophthalmidae, Boreomy- sinae and ? Stygiomysidae. (iii) Statocyst absent : Lophogastrida, Stygiomysidae and Petalophthalmidae. (iv) Palp present on maxillula : Gnathophausia only (Lophogastrida). (v) Exopodite of maxilliped (first thoracic limb) a small unsegmented lamella or absent : Lophogastrida, Stygiomysidae and Petalophthalmidae. (vi) Second thoracic appendage modified as a gnathopod : most Lopho- gastrida and Mysida in the Stygiomysidae and Eucopiidae the next two pairs are similarly modified. (vii) Endopodites of thoracic appendages 3-8 (or 5-8) without subdivision of the penultimate segments : Lophogastrida, Stygiomysidae, Petalophthal- midae and a few Mysidae. (viii) Pleopods reduced, non-natatory, but biramous in both sexes ; second pair slightly modified in ; Stygiomysidae (see p. 303). (ix) Transverse lamellae arising from the posterior sternal margins of abdominal somites 3, 4 and 5 ; Stygiomysidae, though here the protopodites of the pleopods are also incorporated in the lamellae. (Text-figs. 20 and 8 and Caroli, 1937, fig. 3). (x) A short but distinct backward prolongation of the protopodite of the uropod : Stygiomysidae, here it is very much longer (Text-figs. 21 and 13, p). (xi) Exopodite of second thoracic limb normal, multiarticulate : most Mysidacea (except Petalophthalmus, where it is absent). It is long and natatory as in forms with reduced pleopods. The basal segment of this and the following exopodites are long and narrow as in Stygiomysis and Lophogastrida, not blade-like as in most Mysida. Certain characters mentioned by Page in his diagnosis of the family Lepidopidae are of less importance in trying to elucidate affinities since they may vary within one and the same family : (a) Endites or lobes on the first thoracic endopodite may be large, poorly developed or absent, depending on the feeding habits. They are well developed in all three genera of Lepidopidae, absent in Stygiomysidae ; in the family Mysidae they are large in Mysis, one only is present in Mysidetes and they are absent in Mysidopsis. In the Petalophthalmidae they are absent in Hansenomysis, present in Petalophthalmus. In the Lophogastridae they are present in Gnathophausia (Text-figs. 33-34), absent in Lophogaster. (b) Presence or absence of a suture line (diaerisis) on the exopodite of the uropod. In the Lepidopidae there is one in Lepidops (which only shows up well in reflected light) ; it is said to be present in Typhlo- lepidomysis, absent in Spdaeomysis. It is present in the majority of the Lopho- gastrida and in Petalophthalmidae, is faint and incomplete in Siriellinae, absent in other subfamilies of Mysidae and in Stygiomysidae. Rholapophthalmus is excep- tional in having a suture in both exopodite and endopodite. 308 ON A STYGIOMYSIS FROM THE WEST INDIES From the above general survey of the characters used to subdivide the order Mysidacea, it is obvious that, as far as external morphology is concerned, the only constant character separating the Lophogastrida from the Mysida, Lepidopidae and Stygiomysidae, is the presence of podobranchiae or gills on some or all of the thoracic limbs. A negative character like the absence of gills does not necessarily imply that the Lepidopidae and the Stygiomysidae belong to the Mysida as Caroli 1924, p. 513 implies, loss of gills may have occurred more than once in the course of evolution. Several of the Lophogastrid characters are shared by the Lepidopidae, Stygiomysidae and Petalophthalmidae and the large number of oostegites is shared by the Boreomysinae also (as are natatory pleopods in the male). But I do not see that anything is to be gained by erecting a third suborder for the Lepidopidae and Stygiomysidae. In that case something would have to be done about the Petal- ophthalmidae and perhaps the Boreomysinae as well. In the meantime, I prefer to leave both families (Lepidopidae and Stygiomysidae) in the Mysida since podo- branchiae are absent and the pleopods are reduced. These two families are very closely related to each other and differ in several important characters from the rest of the Mysida the similar pleopods, the transverse lamellae on abdominal somites 3, 4 and 5 and the greater or lesser prolongation of the protopodite of the uropod. Their relationship to each is much the same as that of the Lophogastridae and Eucopiidae. Whatever future specialists may decide to do with the Lepidopidae, I am convinced that the Stygiomysidae should be ranked along with it even if there should be fewer than seven pairs of oostegites in Stygiomysis. The family Lepidopidae requires redefinition, because certain characters were overlooked at first ; also for comparison with that of the Stygiomysidae, which is the same as for the genus Stygiomysis (see p. 286). DIAGNOSIS OF THE FAMILY LEPIDOPIDAE : Cavernicolous Mysidacea exhibiting the " caridoid fades'". Carapace relatively long, smooth or with a sinuous furrow ; produced antero-laterally beyond the rounded rostral margin, postero-lateral " wings " well developed, reaching the first abdominal somite laterally, but leaving the last two thoracic tergites exposed dorsally. (A scale-like process on penultimate thoracic tergite in Typhlolepidomysis) 1 . Abdomen, including telson, half as long again as the cephalothorax, not broader in middle, sixth somite as long as, or longer than, fifth. Telson truncate or tapering ; spinules or spines on distal half of each lateral margin as well as on the rounded apex (spinulation not mentioned in Spdaeo- mysis). Ocular scales quadrangular, distal margin oblique, without or with traces of visual elements (a few ommatidia in Lepidops). Antennular and antennal flagella as long as (Lepidops), or far exceeding, the length of the body (Typhlolepidomysis) 1 ; antennal scale well developed, with setae on inner and outer margins. Labrum oval, longer than wide (Lepidops}. Labium with paragnatha well separated proxi- mally, but not as long as in Stygiomysidae. Mandible with well developed incisor and molar processes ; a row of strong thorny spines on second and third segments of the palp. Maxillula small, with a 2-segmented reflexed palp ending in a long curved 1 Since this paper went to press I have collected specimens of Spelaeomysis bottazzii from the type locality. It has enormously long antennae and a scale behind the dorsal posterior emargination of the carapace. Typhlolepidomysis is therefore a synonym of Spelaeomysis. 24.xi.5Q. ON A STYGIOMYSIS FROM THE WEST INDIES 309 seta ; all apical spines on the distal endite of about same size. Maxilla with setae, some of which are distally plumose ; segments of endopodite quite long. Maxilliped not pediform, with well developed endites and the penultimate segment broadened ; exopodite a small unsegmented lamella with or without fine marginal setules ; epipodite rather long and narrow. Second thoracic appendage with exopodite longer than endopodite, which may be modified as a gnathopod. Thoracic endopo- dites 3 to 8 all slender ambulatories. Penultimate segments of thoracic endopodites not subdivided. Pleopods alike in both sexes, apart from the slight sexual dimorphism exhibited by the second pair ; they increase in size from the first to the fifth pair, are biramous, with one-segmented endopodite and three-segmented exopodite (four-segmented in the second pair in the female). A short but distinct transverse chitinous lamella projects from the posterior sternal margin of abdominal somites 3, 4 and 5, with which the protopodites of the pleopods are not incorporated. Uropod with a small but distinct backward prolongation of the protopodite, armed with a terminal spine (? overlooked in Typhlolepidomysis) ; statocyst absent ; a sinuous suture line on the exopodite (absent in Spelaeomysis). Seven pairs of oostegites. Type genus : Lepidops (Fage) 1924, 1925. L. servatus (Fage) is stated to be planktonic : " il existait dans ce petit lac de Machumwi-Ndogo un veritable plancton, compose de petits Schizopodes a yeux tres reduits. Us ne se sont pas fait prendre aux nasses, mais nous en avons receuilli un grand nombre pres de la surface, au filet fin ". Allaud et Jeannel in Jeannel and Racovitza (1914, p. 381). III. NOTE ON THE PARAGNATHA IN SOME MYSIDACEA As a rule, little or no attention is given to the labium or lower lip when the mouth- parts of Malacostraca are examined unless, of course, the feeding habits are being studied in detail. It is scarcely surprising, therefore, that in the subject index of the Ray Society Monograph of the British Mysidacea (Tattersall & Tattersall, 1951) the labrum or upper lip is included, but the labium is not mentioned under any of its synonyms (hypostoma, metastoma, paragnatha). In Stygiomysis, however, the paragnatha are so conspicuous that they could not possibly be overlooked (Text-figs, i and 3, / and Text-fig. 23, p) ; yet Caroli does not mention them in his description of S. hydruntina (1937). Michael Sars (1862, pi. I, figs. 25 and 26) gave two minute sketches of the labium of Lophogaster typicus and G. O. Sars not infrequently figured the posterior, as well as the anterior lip. In his Challenger Report for example, he included small figures of the labium of Gnathophausia and Eucopia (Lophogastrida), of Anchialus and Euchaetomera (Mysida) (Sars, 1885, pi. 8, fig. 5 ; pi. 9, fig. 8 ; pi. 34, fig. u ; pi. 35, fig. 7 and pi. 37, fig. 5). These suggest that the labium is more deeply cleft in the Mysida than in Lophogastrida. The paragnatha are also figured in situ for Hemimysis by Cannon and Manton (1927, pi. in, fig. 4) and for Lophogaster by Manton (1928, pi. in, fig. 6) ; and there are others. But I do not recall having seen any illustration of the paragnatha quite like those of Stygiomysis. In lateral aspect, they are unusu- ally pendant, falling well below the level of the labrum, and the distal part of each 3io ON A STYGIOMYSIS FROM THE WEST INDIES 0.5 mm ON A STYGIOMYSIS FROM THE WEST INDIES 311 is spoon-shaped. (Text-fig, i, /). In ventral aspect, the paragnatha are widely separated proximally, and the spoon-shaped parts conceal much of the mandibles and may over-reach the posterior rim of the labrum (Text-fig. 3, / and Text-fig. 23, p). I am not sure whether there is a subsidiary basal lobule or whether it is just a twist to form a sort of neck (n) or stalk for the spoon-like head. I have examined the paragnatha in several genera of Mysidacea for comparison with those of Stygio- mysis. But there is far more variation than I had anticipated and time is lacking for a detailed comparative study. Nouvel (1943) made a very careful study of the labrum in most of the species that he described and figured in his excellent Report. But, as a rule, he does not mention the labium except in Thalassomysis tatter salli in which he says that it is " tres aberrant. II est aussi tres proeminent et asymetrique. Sa partie droite distale [right paragnath] comporte une cavite ou vient se loger la dent du labrum. Sa partie gauche epouse aussi le contour du bord posterieur gauche du labrum. Presque toute la partie distale est couverte de petites soies analogues a celles du bord gauche du labrum " (p. 59 ; pi. in, fig. 89). But he does not say what he considers the most typical form, or forms, of the labium in the Mysidacea. In Eucopia sculpticauda Faxon (Eucopiidae-Lophogastrida) the labium, repre- sented in Text-fig. 32, is not quite as deeply cleft as that of Lophogaster (Lophogastri- dae) as figured by M. Sars (1862, pi. i, fig. 25) ; in the micro-preparation the parag- natha are slightly displaced, owing to pressure of the cover slip, so that a secondary lobule, armed with marginal spinules, is visible (cf. Lophogaster in Manton, 1928, pi. in, fig. 6). Of the highly specialized yet primitive family Petalophthalmidae (Mysida) I have examined Petalophthalmus armiger W. Suhm and Hansenomysis fyllae (Hansen). In Petalophthalmus the labium is deeply cleft (Text-fig. 26, right-hand drawing) ; in ventral aspect, the paragnatha are much longer than wide, are separated proxi- mally by a narrow gap and meet distally. In lateral aspect, each paragnath is deeply concave, almost bowl-shaped, and the cutting edge of the mandible lies in the con- cavity (Text-fig. 26, left sketch, / and m) . In Hansenomysis the paragnatha are also longer than broad and in situ are not widely separated proximally (Text-fig. 29, upper sketch, made before mounting on a micro-slide). In the micro-preparation the paragnatha are displaced owing to pressure of the cover-slip and the inner lobules with their marginal spinules are visible (cf. Text-figs. 29 and 32). In Text-fig. 22 the labium of Boreomysis rostrata Illig (Boreomysinae-Mysida) is represented in three different aspects ; the paragnatha are slightly asymmetrical and there is no distinct subsidiary lobule on either of them. An inner knob or boss (a) can be seen in situ between the paragnatha (upper right sketch which is the aspect FIGS. 22-24. Boreomysis rostrata Illig. Fig. 22. Labium or lower lip, in three different aspects. Stygiomysis holthuisi (Gordon). Fig. 23. Right half of labium, in ventral aspect (incisor edge of left mandible also shown) . Fig. 24. Anterior half of male, in dorsal aspect, (first abdominal somite represented by broken line), a. Boss on labium. n. Neck. p. Paragnath. 5 and 8. Fifth and eighth thoracic somites. 312 ON A STYGIOMYSIS FROM THE WEST INDIES 0.5 FIGS. 25-28. Petalophthalmus armiger W. Suhm. (Reg. no. i94O,viii.5. 568-569). Fig. 25. Right maxillula. Fig. 26. Labrum, right mandible, paragnath and maxillula, in lateral aspect ; also right paragnath in ventral aspect. Lepidops servatus (Fage). Syntype from Paris Museum. Fig. 27. Labrum, labium, mandibles and maxillulae, in ventral aspect. Fig. 28. Same, in lateral aspect, at higher magnification. /. Parag- nath of labium. m. Mandible, mx. Maxillula. p. Mandibular palp. u. Labrum or upper lip. ON A STYGIOMYSIS FROM THE WEST INDIES 313 from which most other paragnatha are figured). In the micro-preparation of Indo- mysis annandalei Tattersall (Mysini-Mysida) the labrum and labium are somewhat flattened by the cover-slip (Text-fig. 30, upper sketch) ; the labium is more deeply cleft than in Eucopia (Text-fig. 32) ; each paragnath is almost as wide as long and may touch the posterior margin of the labrum ; there is apparently no secondary lobule (upper and lower sketches) . Of Lepidops servatus (Fage), which is a very minute and fragile species, I have examined a damaged specimen in which the paragnatha are flexed away from the mandibles and the right maxillula is somewhat displaced. The camera lucida sketches were prepared before any attempt was made to mount the fragment (Text-figs. 27 and 28). The paragnatha are obviously rather widely separated proximally and, of all those examined, they approach most nearly to those of Stygiomysis although they are smaller and have no " neck " portion (or subsidiary lobule see p. 311). In my preliminary note on " Rhopalonurus " holthuisi I referred to the paragnath or main lobe as a lappet (Gordon, 1958, p. 152) but it probably is not comparable with the small moveably articulated lappets found in some Tanaidacea. The proxi- mal portion (Text-fig. 23, n) may represent the subsidiary lobule which is present in Eucopia and in Hansenomysis (Text-figs. 32 and 29 in each the lobule is beset with spinules) or it may just be the result of a slight twist to form a " neck " or stalk. The unusual form and size of the paragnatha in Stygiomysis is probably a secondary specialization connected with its feeding habits, as are the other peculiarities of the mouthparts. The large maxillula with the remarkably enlarged spine on the distal endite has already been referred to ; I have figured the maxillula of most of the species of Mysidacea examined for comparison with that of Stygiomysis (cf. Text- figs 25-28 and 29, 30, 32 with Text-figs. 3 and 5). When the anterior (uppermost) spine on the distal endite is larger than the others this difference in size is not very pronounced (Text-figs. 25, 32 and 30). The feeding habits have been studied in detail only in a few species of Mysidacea although they have been deduced or inferred from an examination of the mouthparts in other species. But the significance of the marked differences in both form and size of the paragnatha, for example, is not known. A summary of our knowledge of food and feeding habits is given in Tattersall & Tattersall, 1951, pp. 29-33 and other references are listed in their subject index under feeding habits, food, filter feeding and raptorial feeding. Many Mysids employ the filtratory method and the raptatory method quite independently, e.g. Hemimysis (Cannon and Manton, 1927) and Gnatho- phausia, although in the latter there can be no auxilliary food current created by the thoracic exopodites as in other primitive filtratory Malacostraca (see Manton, 1928, p. 117). Others are incapable of filter feeding and depend entirely on large food masses, like Lophogaster (Manton, 1928, p. no) and Petalophthalmus (Tattersall & Tattersall, 1951, p. 113). The number of thoracic limbs adapted to aid in feeding varies considerably. In the Lepidopidae, as in many Mysids, the first and second pairs are modified as maxilliped and gnathopod respectively. But others may also assist, though not specially modified e.g. in many Mysidae all the thoracic endopodites may assist in dealing with large food masses, although this is not their main function (Tattersall & Tattersall, 1951, p. 30). In Stygiomysis the first four ON A STYGIOMYSIS FROM THE WEST INDIES V 01 ' 0.5 5 FIGS. 29-32. Hansenomysis fyllae (Hansen). (Reg. no. i9og.xii.24. 100-103) Fig. 29. Labium and maxillula, in ventral aspect (paragnatha displaced by pressure of cover slip) ; above, paragnatha figured before mounting the preparation. Indomysis annandalei Tattersall. (Reg. no. 1946. xi. 26. 45-64). Fig. 30. Labrurn, labium, left mandible and maxillula ; below, right paragnath. (Preparation slightly flattened by cover slip). Eucopia sculpticauda Faxon. (Reg. no. 1949. vii. 13. 65-75). Fig. 31. Right maxilliped, inner (or dorsal) surface. Fig. 32. Labium and left maxil- lula, in ventral aspect. (Scales for Figs. 29-32 all = 0.5 mm.), b. Sternal surface. ex. Exopodite. /. Paragnath. p. Mandibular palp. u. Upper lip or labrum. 4. Segment of endopodite distal to " knee ". ON A STYGIOMYSJS FROM THE WEST INDIES 315 pairs are modified as maxilliped and gnathopods ; the same is true of the Eucopiidae, but here the following three pairs (5-7) are long and prehensile (see pp. 305 and 319). This brings up the question of the terminology of the thoracic appendages, which is discussed in the next section. IV. TERMINOLOGY OF THE THORACIC APPENDAGES. (a) TERMINOLOGY OF THE APPENDAGES. Various names have been given to the thoracic appendages of Malacostraca according to the special functions for which they may be adapted. Dr. O. S. Tattersall writes of the Mysidacea : " since different appendages may be adapted for this or that function in different genera, I consider such terms most confusing, and in this work I propose to use one term, thoracic limbs, for these eight appendages, and to refer to each under its appropriate number, irrespective of the particular function for which it has become adapted." (Tattersall & Tattersall, 1951, pp. 18-19.) That there is some confusion is apparent if one examines certain statements in the joint work just cited. For example, under the definition of the order Mysidacea (pp. 11-12) " first, and sometimes the second, pairs modified as gnathopods or maxillipeds " might be interpreted in different ways that both pairs are either gnathopods or maxillipeds, or that one pair is to be regarded as maxillipeds the other as gnathopods. On p. 24 " second thoracic limb (= gnathopod or 2nd maxilliped) " is also ambiguous are the terms used synonymously or not ? On p. 97 " endopods of the second to the fourth pairs of thoracic limbs . . . developed as gnathopods " implies that the first pair are maxillipeds. Most of the confusion centres round these two terms. Sometimes they are used synonymously ; sometimes each is used in a narrower sense, the anterior pair (or pairs) which are more markedly modified for " Nahrungszerkleine- rung " being called maxillipeds in contrast to the less markedly modified succeeding pair (or pairs) of gnathopods (see Zimmer, 1927, pp. 302-303). Nor is there agreement as to the number of pairs of thoracic appendages which are modified as maxillipeds in the various orders of Malacostraca (the term gnathopod is commonly used in literature dealing with the superorder Peracardia, rarely if at all in that dealing with Eucarida, Hoplocarida or Syncarida). In primitive types none of the thoracic limbs are differentiated as maxillipeds (Caiman, 1909, p. 145) ; this is the case in Nebaliacea and Euphausiacea according to most authors. In other orders one, two, or three pairs are generally regarded as maxillipeds. With regard to the Stomatopoda Caiman (1909, p. 322) wrote : " The first five pairs of thoracic appendages are similar in structure and are commonly called maxillipeds, though, as they possess no endites or other adaptations for mastication, the name is hardly appropriate ", Yet it is still often said that the Stomatopods have five pairs of maxillipeds here the term gnathopod might be more appropriate. Hansen (1925, pp. 94, 128 and 154) thought that throughout the whole of the Malacostraca there is one pair of real maxillipeds. In his summing up (p. 154) he says : " Maxil- lipeds and thoracic legs. . . . They are similar in Leptostraca, while in the other orders the first pair, the real maxillipeds, differ somewhat or considerably or very much from the following pairs, the legs." Yet even he inadvertently introduces 316 ON A STYGIOMYSIS FROM THE WEST INDIES further confusion because : (i) the reader has to bear in mind that his thoracic legs 1-7 are in reality thoracic appendages 2-8 and (ii) under the Decapoda he has reverted to the usual terminology of maxillipeds 1-3 and peraeopods or walking legs 1-5 (= thoracic limbs 4-8). Also under the Mysidacea he speaks of " inter- mediate " legs. The terminology of the thoracic limbs has recently been discussed by Slewing (1958, pp. 250-251). The maxillipeds, he says, are chiefly adapted for chewing or filtering food and are characterized above all by the presence of endites. " Sind die Endite deutlich grosser als die den nach hinten folgenden Extremitaten, so wird man sie betreffenden Extremitaten als Maxillipeden von den folgenden differenzieren. In diesem Sinne ist das 1 Thorakalbeinpaar, zum Teil auch das zweite, der Mysidacea als Maxilliped zu bezeichnen, denn es besitzt an der Basis einen abgegliederten Endite ; Kaufortsatze werden ausserdem an Praeischium und Ischium beobachtet." Gnathopods, on the other hand, have he says their distal segments modified, often to form chelae or subchelae ; but according to Tattersall & Tattersall (1951, p. 24) the first thoracic limbs of Mysidacea may sometimes be subchelate. Siewing's statement concerning the Mysidacea agrees in part with that of Caiman (1909, p. 174) : " The first pair of thoracic limbs are always specialized as maxillipeds. In the Lophogastridae and Eucopiidae they are without distinct endites (italics mine) . In the Mysidae an endite is generally borne on the basipodite, and sometimes also by each of the two following segments." Tattersall & Tattersall (1951, p. 97) also say that in Eucopia there are " no gnathobasic lobes on second, third or fourth segments ". If Siewing's definitions are strictly applied, the first thoracic limb of Eucopia, though smaller than, and appreciably different from, each of the three succeeding pairs of appendages, would also have to be called a gnathopod (see Text-fig. 31 of this paper) . The same applies to the first thoracic limb of, e.g. Mysidopsis which also has no endites (gnathobasic lobes) on second, third or fourth segments (Tattersall & Tattersall 1951, p. 316) and to that of Stygiomysis although it is much smaller, and more intimately associated with the mouth parts, than are the three following pairs (Text-figs. 7, 9 and 10). Caroli (1937, pp. 22 and 226), who refers to the first pair as maxillipeds, gives the neatest possible definition of the next three pairs with his " Toracopodi 2-4 (gnatopodi) massillipediforme "! Typhlolepidomysis, on the other hand, exemplifies the two categories exceptionally well, having large endites on the first pair which are not pediform, none on the pediform second pair which have strong spines on the more distal segments. Unfortunately, Villalobos (1951, p. 205) refers to the latter as " primer par de pereiopodos " in the text and as " maxilipedo del secundo par " in the legend to Text-fig. 10. In Lepidops the second pair differs but slightly from the following pairs of thoracic limbs. In the family Petalophthalmidae the second thoracic limb has a large lamellar expansion on the ischium (merus of some writers) which, with its partner, forms a floor to the whole mouth region. (I have seen a specimen which had been preserved while enjoying a meal, and these lobes apparently help to retain the food mass while it is being chewed). Should this appendage be termed a maxilliped? It is much larger than the first thoracic limb (which in Hansenomysis has no endites) and its distal segments are modified to assist in manipulating the food mass. It seems illogical ON A STYGIOMYSIS FROM THE WEST INDIES 317 to term the first thoracic appendage a maxilliped in Mysis but not in Mysidopsis, in Petalophthalmus but not in Hansenomysis, in Gnathophausia (Text-figs. 33 and 34) but not in Lophogaster see p. 316. Form depends on function and the feeding, locomotor and other habits of the animals determine how many appendages will be modified, and to what extent. The requirements of a raptorial animal differ FIGS. 33-34. Gnathophausia ingens (Dohrn). (Unregistered specimen from Madeira). Fig. 33. Right maxilliped, in ventral aspect. Fig. 34. Part of same, in dorsal aspect. e. Endite. ep. Epipodite. ex. Exopodite. 3 and 4. Segments of endopodite on either side of " knee " joint. 318 ON A STYGIOMYSIS FROM THE WEST INDIES .82 a Mysidacea and other ERACARIDA vi, Figs. ioe,f. s & $ E- 'S .- x > S _; * E i a a 1 if t/j t/> * A fr V ' c'ST I A If 6 O " CX'S A S| ON A STYGIOMYSIS FROM THE WEST INDIES 319 from those of a filter feeder, even if the latter can also manipulate large food masses. The habitat to which an animal is adapted, even the niche within the habitat, may lead to modifications in both filtratory and raptatory feeders. Eucopia, with its long, slender " prehensile " endopodites 5-7, may remain suspended in mid water and seize any food that happens to come within range of these long legs, passing it on to the short gnathopods (2-4) ; it may even pursue its prey since the pleopods are well developed and natatory (see also p. 305). I fully agree with Siewing that terms ought to be properly defined, and that the terms under discussion need re- defining. But, for the reasons given above, I do not think that one can apply his definitions too rigidly, within the order Mysidacea at any rate. Cannon (1927) has shown that all the thoracic limbs in Nebalia are specially adapted for a unique mode of filter feeding. More recently Barkley (1940) has figured in great detail thoracic limbs 1-6 of Euphausia and has found that these six pairs together form an elaborate basket or filter for straining minute organisms from the water (the last two pairs of thoracic appendages are greatly reduced). In both these genera the similarity of the thoracic limbs would appear to be a second- ary specialization rather than a primitive character. (b) TERMINOLOGY OF THE SEGMENTS OF THE ENDOPODITE : There is also consider- able diversity of opinion regarding the terminology of the segments in the endopodite of the thoracic appendages. In Stygiomysis holthuisi, for example, it is possible to interpret the parts numbered 1-6 in Text-figs. 9, 10 and n in three different ways : (i) ischium, merus, carpus, propodus, dactylus, and claw ; (ii) preischium, ischium, merus, carpus, propodus, and dactylus (with which the claw is incorporated); (iii) preischium, ischium, merus, carpo-propodus, dactylus, and claw. According to (i) the " knee " joint is at the junction of carpus and propodus ; according to (ii) and (iii) it lies between merus and carpus. This is why I have simply numbered the segments of the endopodite those distal to the basis which bears the exopodite and is part of the protopodite. The different terminologies which are in use are best shown in tabular form (Table I, which may not be complete). To the student reared on Caiman's text book terminology A will be familiar ; according to this the " knee " joint is sometimes between merus and carpus, sometimes between carpus and propodus. The other terminologies (B, G and D) we owe to Hansen (1925). He maintains that there is an additional segment, the preischium, which is distinct in Peracarida but is often fused with the ischium (Eucarida) or with the basis (Syncarida). He also maintains that throughout the Eumalacostraca the " knee " is at the junction of merus and carpus and that, where there appears to be only two segments distal to the " knee ", the carpus is fused with the propodus (the claw is not a true segment and is not included). As shown in the table, this fusion of carpus and propodus occurs very frequently in Mysidacea and throughout the other Peracaridan orders. Siewing (1958) has recently studied Thermosbaena and his interpretation of the segmentation is shown as Ei and E2. In minute animals like Thermosbaena and the Bathynellacea there is a reduction in the number of segments and it is by no means easy to decide where the " knee "is. Siewing thinks that here there is no preischium, and that the ischium is either fused with the basis or more or less completely fused with the merus. 320 ON A STYGIOMYSIS FROM THE WEST INDIES In Stygiomysis holthuisi the number of segments in the endopodite is identical in each of the eight pairs of thoracic limbs (Text-figs. 7, 9-11). In each there are three segments proximal to the " knee " and three distal to it (including the claw or claw-like dactylus, according to the terminology used). Hansen would doubtless have regarded segment 4 as a fused carpo-propodus (64) rather than as the carpus followed by the propodus and claw-like dactylus (Bi). In Tattersall & Tattersall (1951, p. 24) it is stated that, in both first and second pairs of thoracic limbs of Mysidacea, carpus and propodus are always fused to form a single segment. Yet in fig. 42, C on p. 199 the first thoracic endopodite of Erythrops elegans (G. O. Sars) is represented as having four segments proximal to the " knee " joint and two plus claw distal to it. This can only be interpreted (if correct and not a slight lapse on the part of the artist) as : preischium, ischium,. merus, carpus, [" knee "], propodus, dactylus and claw. According to Hansen the " knee " joint in Peracarida does not come at the junction of carpus and propodus. Since fig. 42, D shows the second endopodite of Erythrops elegans with only three segments proximal to the " knee ", fig. 42, C may not be correct. Hansen (1925) based his interpretation of the segmenta- tion, in part at least, on a study of the muscles (see his figures of the distal segments of the thoracic endopodite in Mysidacea on pi. vi). In Stygiomysis there appear to be no muscle bands in segment 5, which might support Hansen's interpretation of 5 and 6 being dactylus and claw respectively. Perhaps the time has come for a thorough reappraisal of Hansen's work along the lines of the detailed functional analysis of limb structure undertaken in recent years by Dr. Manton in terrestrial Arthropoda (Manton, 1950-58 a series of papers on the Evolution of Arthropodan Locomotory Mechanisms ; the work is still in progress). V. ADDITIONAL NOTE ON SPELAEOGRIPHUS LEPIDOPS GORDON Order SPELAEOGRIPHACEA Gordon, 1957 Spelaeogriphacea Gordon, 1957, P- 45- Spelaeogriphacea, Slewing, 1958, p. 267 ; Kaestner, 1959, p- 918. " Reihe " Spelaeogriphacea Slewing, 1959, p. no, Fig. 65. SYSTEMATIC POSITION. Slewing (1958, p. 247) placed the order Spelaeogriphacea among those Eumalacostraca with the telson and last abdominal somite fused. But, as stated in my original description and as shown in PI. 4 of this paper, the sixth pleonite is free from, and much larger than, the telson. In a later work, which has just appeared, Siewing (1959, p. no, fig. 65 cited in his 1958 paper as if it had already been published) has raised the Spelaeogriphacea to the rank of a "Reihe" or superorder, without apparently stating his reasons for doing so. Since Spelaeogriphus, the only known genus, undoubtedly possesses a Peracaridan brood pouch composed of oostegites, I still regard the Spelaeogriphacea as an order belonging to the superorder (" Division " of Caiman) Peracarida. Its precise relation- ship to the other Peracaridan orders cannot be determined until the internal anatomy and the embryology of Spelaeogriphus are known. However, from its external ON A STYGIOMYSIS FROM THE WEST INDIES 321 morphology, it appears to have affinities with the Tanaidacea, since it has a small but distinct carapace and a pair of cup-shaped respiratory organs beneath the cara- pace, near the bases of the first thoracic limbs (the maxillipeds). Siewing (1956, p. 87) has pointed out that in the Tanaidacea " der hinter dem Maxillipeden an der Korperwand liegende loffelartige Fortsatz ist kein Epipodit, wie Dohrn bei seinen embryologischen Untersuchungen zeigen konnte." (Dohrn, 1870, pi. xi). The respiratory organ in Spdaeogriphus may also be of this kind and not a true epipodite, but I called it an epipodite since that is the term commonly used in systematic descriptions of Tanaidacea. However, I could not include Spelaeogriphus in the order Tanaidacea, since it differs in many respects from that group, having certain affinities with the Syncarida and with Monodella and some which are unique. In his earlier paper Siewing (1957, p. 267, phylogenetic diagram) placed it within the Peracarida, provisionally near the Isopoda-Tanaidacea line. Unless Dr. Siewing has studied material of Spelaeo- griphus lepidops and has found that it differs markedly, anatomically and /or embryologically, from the other Peracaridan orders, his decision to separate Spelaeo- griphacea from the " Reihe " Peracarida is premature, and in my opinion this step is unnecessary. It is a pity that this alteration was made in a text book because in all probability his phylogenetic scheme will be reproduced in future text books. The one thing that all Peracaridan orders have in common is that the young are incubated until a late stage of development in a ventral thoracic brood pouch or marsupium ; Siewing's (1958) separation of the order Thermosbaenacea from the true Peracarida is much more reasonable, since in both the known genera there is a dorsal carapacial marsupium. Stella (1959, p. 122) has accepted the " Reihe " Pancarida although she unfortunately refers to it as a " phylum " . Raising the Spelae- ogriphacea to superordinal rank does not solve the problem of its relationships. The greatest caution should be exercised in establishing higher categories like superorders and subclasses until the animals have been thoroughly studied. Genus SPELAEOGRIPHUS Gordon, 1957 Spelaeogriphus Gordon, 1957, P- 3 2 - Spelaeogriphus, Kaestner, .1959, p. 918. Spelaeogriphus lepidops Gordon, 1957 Spelaeogriphus lepidops Gordon, 1957, P- 33> Text-figs. 1-26. Spelaeogriphus lepidops, Kaestner, 1959, p. 918, Text-fig. 816. ADDITIONAL MATERIAL. Bats Cave, Table Mountain, 3o.ix.56, six specimens fixed in Bouin's solution (4 <$, i ovigerous and i smaller $). From a stream at bottom of Bats Cave (temp, of water 50-5 F.), i.ix.57, seven specimens (5 $, i ovigerous $ and 2 immature specimens). REMARKS. The specimens were collected by the South African Spelaeological 322 ON A STYGIOMYSIS FROM THE WEST INDIES Association ; the 1957 material was fixed in 96 per cent ethyl alcohol (97 parts) and glacial acetic acid (3 parts). This fixative seems to be too drastic, causing the speci- mens to curl up to such an extent that in one instance the uropods got entangled among the thoracic appendages bunched up beneath the carapace region. I have had several requests for a good figure of the entire animal, suitable for reproduction in a text book. My original Text-fig, i was incomplete, since I did not include the appendages, and Kaestner (1959, fig. 816) has based his figure on Dr. Barnard's sketch which I reproduced as Text-fig. 26 of my 1957 paper. The anten- nulae and antennae are not represented and these alter the look of an animal appre- ciably ; in Kaestner's reproduction the first three thoracic exopodites look not unlike the gills of Lophogastrid Mysids. I take this opportunity, therefore, of publishing an original drawing of the whole animal, an ovigerous female, in dorso-lateral aspect as PI. 4. This figure will also illustrate my article on Spelaeogriphacea which will appear in due course in the McGraw-Hill Encyclopedia of Science and Tech- nology. The female represented in PI. 4 measures 6 mm. in length, from rostral apex to tip of telson, and the antenna is almost or quite as long as the body. The two antennular flagella are slightly unequal (they are more so in the male) ; the inner flagellum has about 30, the outer one about 24, segments. The faint line down the right side of the body, from carapace to first abdominal somite, is an artefact, a slight fold or crease in the integument. The marsupium contains ten to twelve rather large ova. I have not attempted to straighten out the thoracic limbs, which in preserved material tend to bend inwards beneath the body, because they easily break off if touched. There is not much to add to my original description, since the animals do not seem to grow much larger ; the largest male measures 8-6 mm. in length. Its antenna is at least 8 mm. long ; the inner flagellum of the antennula is appreciably longer than the outer one and the number of segments in each is about 45 and 26 respectively. In my original description I used Caiman's terminology for the various segments of the thoracic limbs (see Table I, A). But, according to Hansen (1925), there is always a distinct preischium in the Peracarida (Table I, Bi to 64). The segments of the endopodite would thus be termed : preischium, ischium, merus, [" knee " joint], carpo-propodus and dactylus (Table I, 64). I have re-examined these appendages and in some there seems to be a faint indication of separation of the claw from the dactylus. In one specimen the distal sharp point is broken off neatly, beyond the longest spine on the ventral margin ; this part doubtless represents the claw or nail (Gordon, 1957, Text-figs. 19 and 21-25). The entire length of the gut is filled with a dark brown or blackish mass, suggesting that the animal feeds largely on detritus. The registration numbers of the type specimens in the British Museum Collection are : holotype (^), 1957. xi. 12.1 and allotype ($), 1957. xi. 12. 2. I have a correction to make : in my 1957 paper (p. 43) I said that the Tanaidacea are entirely marine. But I overlooked a paper by Stephensen (1936) in which he described Tanais stanfordi from a freshwater lake in the Kurile Islands and mentions at least one other freshwater species from South America. ON A STYGIOMYSIS FROM THE WEST INDIES 323 LIST OF WORKS REFERRED TO. ALLAUD, CH. & JEANNEL, R. 1914. In Jeannel, R. and Racovitza, E. G. Enumeration des grottes visitees 1911-1913. (5) 413. Arch. Zool. exp. gen. (Biospeologica xxxiii) 53 (7) : 325-558, 5 text-figs, [p. 381]. BARKLEY, E. 1940. Nahrung und Filterapparat des Walkrebschen Euphausia superba Dana. Z. f. Fisch. Beiheft i, Walforschung i, Enhalt IV : 65-156, 39 text-figs. BOLIVAR Y PIELTAIN, C. 1943-44. Exploracion biologica de algunas cavernas de Cuba. Ciencia Mexico. 4 (11-12) : 301-304. 1950. Estudo de una Civolana cavernicola nueva de la region de Vallas, san Luis Potosi, Mexico. Ciencia Mexico 10 (7-8) : 211-218. CALMAN, W. T. 1909. A Treatise on Zoology, edited by Sir Ray Lankester, Part 7, Appendi- culata, Fasc. 3 Crustacea. London (A. & C. Black) : viii + 346 pp. Text-figs. - 1932. A cave-dwelling Crustacean of the family Mysidae from the Island of Lanzarote. Ann. Mag. nat. Hist. (10) 10 : 127-131, 6 text-figs. CANNON, H. G. 1927. " On the feeding mechanism of Nebalia bipes." Trans. R. Soc. Edinburgh. 55 ' 355-37. 7 text-figs. CANNON, H. G. & MANTON, S. M. 1927. On the feeding mechanism of a Mysid Crustacean, Hemimysis lamornae. Trans. R. Soc. Edinburgh, 55 : 219-253, 4 pis., 16 text-figs. CAROLI, E. 1924. Su di un misidaceo cavernicolo (Spelaeomysis bottazzii n.g., n.sp.) di terra d'Otranto. Atti Ace. Naz. Lincei Roma, 33 (2) : 512-513. - 1937. Stygiomysis hydruntina n.g., n.sp., Misidaceo cavernicolo di Terra d'Otranto, rappresentante di una nuova famiglia. Nota preliminare. Boll. Zool. Torino, 8 : 219-227, 3 text-figs. CHAPPUIS, P. A. 1931. Copepoda Harpacticoidea der Deutschen limnologischen Sunda- Expedition. Arch. Hydrobiol. Stuttgart. Suppl. 8 : 512-584, 152 text-figs. GREASER, E. P. 1936. Crustaceans from Yucatan. Publ. Carnegie Inst. Washington, 457 : 117-132. 43 text-figs. 1938. Larger cave Crustacea of the Yucatan Peninsula. Publ. Carnegie Inst. Washington, 491 : 159-164, 8 text-figs. DOHRN, A. 1870. Untersuchungen iiber Bau und Entwicklung der Arthropoden. VII. Zur Kenntnis vom Bau und der Entwichlung von Tanais. Jenaische Zeitschr. Nat. 5 : 293-306, pis. ii and 12. FACE, L. 1924. Sur un type nouveau de Mysidace des eaux souterraines de 1'ile de Zanzibar. C.R. Acad. sci. Paris. 178 : 2127. - 1925. Lepidophthalmus servatus Fage. Type nouveau de Mysidace des eaux souterraines de Zanzibar. (Biospeologica 51). Arch. zool. Exper. Gen. 63 : 525-532, 7 text-figs. 1941. Mysidacea : Lophogastridae I. Dana Rept. Copenhagen, 19 : 1-52, 51 text-figs. 1942. Mysidacea : Lophogastridae II. Dana Rept. Copenhagen, 23 : 1-67, 42 text-figs. GORDON, I. 1957. O n Spelaeogriphus, a new cavernicolous Crustacean from South Africa. Bull. Brit. Mus. (Nat. Hist.), Zool. 5 (2) : 31-47, 26 Text-figs. 1958. A new subterranean Crustacean from the West Indies. Nature, Land. 181 : 1552-1553, 2 text-figs. HANSEN, H. J. 1925. Studies on Arthropoda. Copenhagen. Vol. II, 176 pp., 8 pis. KAESTNER, A. 1959. Lehrbuch der Speziellen Zoologie. Pt. I : Wirbellose, Lief. 4. Jena. pp. 659-979, text-figs. 630-861. MANTON, S. M. 1928. On some points in the anatomy and habits of the Lophogastrid Crustacea. Trans. R. Soc. Edinburgh, 56 : 103-119, 3 pis., 2 text-figs. 1950-1958. The Evolution of Arthropodan Locomotory Mechanisms. Parts 1-6. /. Linn. Soc. (Zool.) London, 41 : 529-570, 4 pis. 8 text-figs.; 42 : 93-117, 5 text-figs.; 42 : 118-166, 12 text-figs., 6 pis.; 42 : 299-368, 4 pis., 8 text-figs.; 43 : 153-187, i pi., 8 text-figs.; 43 : 487-556, i pi., 31 text-figs. MONOD, TH. 1940. Thermosbaenacea in Bronn's Klassen und Ordnungen des Tierreichs, 5, i (4) : 1-24, 32 text-figs. Leipzig. 324 ON A STYGIOMYSIS FROM THE WEST INDIES NICHOLLS, G. E. 193 1 . Micraspides calmani, a new Syncaridan from the West Coast of Tasmania. /. Linn. Soc. Lond. 37 : 473-488, 2 pis., i text-fig. NOUVEL, H. 1943. Mysidaces provenant des campagnes du Prince Albert xer de Monaco. Res. Camp. sci. Monaco, 105 : 1-128, 180 text-figs., 5 pis. SARS, G. O. 1885. Report on the Schizopoda collected by H.M.S. Challenger during the years 1873-76. Zool. Challenger Exp. 13, Pt. 37 : 228 pp., 38 pis., text-figs. SARS, M. 1862. Beskrivelse over Lophogaster typicus, en maerkvaerdig form af de lavere tifoddede Krebsdyr . . . Universitets-program . . . pp. iv and 37, 3 pis. 4 Christiania [With resume in French]. SIEWING, R. 1958. Anatomic und Histologie von Thermosbaena mirabilis. Ein Beitrag zur Phylogenie der Reihe Pancarida (Thermosbaenacea) . A had. Wiss. und Liter atur. Abh. Math. -Nat. Kl. Jahr. 1957 No. 7 : 197-270, 43 text-figs. ~ J 959- Syncarida in Bronns Klassen und Ordnungen des Tierreichs. V, i, 4 (2) : 1121, 65 text-figs. STAMMER, H. J. 1933. Einige seltener oder neue Hohlentiere. Verh. Deutsch. zool. Ges. Leipzig. 35 : 263-266. 1936. Ein neuer Hohlenschizopode, Troglomysis vjetrenicensis n.g., n.sp. . . . Zool. Jahrb. Jena (Syst.) 68 : 53-104, 30 text-figs. STELLA, E. 1959. Ulteriori osservazioni sulla riproduzione e lo sviluppo di Monodella argentarii (Pancarida, Thermosbaenacea). Rivista di Biologia 51 (i) [n.s. vol. xi] : 121-144, 6 text- figs., 2 pis., i table. STEPHENSEN, K. 1936. A Tanaid (Tanais stanfordi Richardson) found in fresh water in the Kurile Islands . . . Annot. Zool. Japan. 15 (3) : 361-373, 5 Text-figs. TATTERS ALL, O. S. 1957. Report on a small collection of Mysidacea from the Sierra Leone estuary, together with a survey of the genus Rhopalophthalmus Illig and a description of a new species of Tenagomysis from Lagos, Nigeria. Proc. zool. Soc. Lond. 129 : 81-128, 15 text-figs. TATTERSALL, W. M. 1939. The Euphausiacea and Mysidacea of the John Murray Expedition to the Indian Ocean. Sci. Rep. John Murray Exped. London, 5 : 203-246, 21 text-figs. 1951. A review of the Mysidacea of the United States National Museum. Bull. U.S. nat. Mus. 201 : 292 pp., 103 text-figs. TATTERSALL, W. M. & TATTERSALL, O. S. 1951. The British Mysidacea. Ray Society London 136 : pp. viii + 460, frontispiece and 118 text-figs. VILLALOBOS, A. 1951. Un nuevo misidaceo de las Grutas de Quintero en el Estado de Tamaulipas. An. Inst. Biol. Mexico, 22 : 191-128, 14 text-figs. ZIMMER, C. 1927. Allgemeine Einleitung in die Naturgeschichte der Crustacea (pp. 277-304, text-figs. 216-242) ; Mysidacea (pp. 607-650, text-figs. 592-645) in : Kukenthal und Krumbach, " Handbuch der Zoologie " 3 (i). Berlin (de Gruyter) 4. PLATE 3 Siygiomysis holthuisi (Gordon). Holotype, in dorsal aspect, a. Biramous antennula. b. Antenna, o. Ocular scale. 5. Fifth thoracic somite, i and 6. First and sixth abdominal somites. PLATE 4 Spelaeogriphus lepidops Gordon. Ovigerous female in dorso-lateral aspect, a. Biramous an- tennula. b. Antennal flagellum. g. Respiratory exopodites or gills, o. Ocular scale, s. Antenna! scale or exopodite. 5. Fifth thoracic somite, i and 6. First and sixth abdominal somites. Bull. B.M. (N.H.] Zoo/. 6, 5 PLATE 3 V Bull. B.M. (N.H.) Zool. 6, 5 PLATE 4 CONTRIBUTIONS TOWARD A CLASSIFICATION OF MODERN ISOSPONDYLOUS FISHES WILLIAM A. GOSLINE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 6 LONDON: 1960 CONTRIBUTIONS TOWARD A CLASSIFICATION OF MODERN ISOSPONDYLOUS FISHES BY WILLIAM A. GOSLINE Pp- 325-365 ; 15 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 6 LONDON : 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series corresponding to the Departments of the Museum, and an Historical Series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 6 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued April, 1960 Price Fifteen Shillings CONTRIBUTIONS TOWARD A CLASSIFICATION OF MODERN ISOSPONDYLOUS FISHES By WILLIAM A. GOSLINE THAT the classification of the modern isospondylous fishes (Order Clupeiformes) is unsatisfactory would be admitted by all who have worked with them. The present contributions to the subject are divided into four sections. The first traces one structural complex, the caudal skeleton, through the whole group. The second and third have to do with controversial sections of isospondylous classification. In the final section a revised arrangement of modern forms, down to superfamily, is presented and its rationale discussed. The sections are in logical sequence, but it may also be noted that they are arranged in order of descending assurance on the part of the author. More, perhaps, than usual with classifications, that presented in section four is merely a " progress report ". Regarding nomenclature, the ordinal, subordinal, and superfamily names are formed according to the system proposed by Berg (1940 : 353). The following procedures concerning common names have been adopted. Where an old literature name is well known it will be used. Thus the members of the order Clupeiformes will be called the isospondylous fishes, and those of the suborder Esocoidei the haplo- mous fishes. Where a common name in this paper ends in oid, it refers to a member of a suborder. For example, the term salmonoid fishes will be used to designate all of the forms of the Suborder Salmonoidei and not just those of the superfamily Salmonoidae. No common names will be used for superfamily 1 . Bone nomenclature is that of Devillers (1958). The work on which this paper is based was nearly all done at the British Museum (Natural History) and I wish to thank the members of the fish division of that Museum for the facilities extended me and for permission to examine the skeleton and fish collection. I also wish to thank them, notably N. B. Marshall, for the discus- sion and suggestion of various points incorporated into the paper. The following individuals have been kind enough to read part or all of the manuscript and to offer constructive criticism : E. H. Ahlstrom, H. B. Bigelow, D. M. Cohen, D. H. Dunkle, N. B. Marshall, and G. W. Mead. Finally, I am greatly obligated to the John Simon Guggenheim Foundation for the fellowship that has made this work possible. 1 If Berg's termination oidei is used for suborder, the problem of devising different common names for subordinal and superfamily categories is equally great whether the superfamily ending oidea, recom- mended by the London Colloquium, or Berg's oidae is used. ZOOX. 6, 6. 2I 328 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES SECTION i. THE CAUDAL SKELETON OF MODERN ISOSPONDYLOUS FISHES INTRODUCTION The caudal skeleton of teleostean fishes has been the subject of papers by Barring- ton (1937), Blanco (1935), Hollister (1936, etc.), Kolliker (1860), Lotz (1864), Regan (igioa and b), Totton (1914), Whitehouse (igioa and b, 1918), and others. Never- theless, the accumulated information is rather amorphous, and the caudal skeleton is usually either ignored or briefly dismissed in taxonomic papers. A large part of the difficulty would seem to stem from the failure to establish stable reference points as a basis for comparing one caudal skeleton with another : hence the intro- ductory remarks that follow. In a relatively simple caudal skeleton like that of Amia (Goodrich, 1909, fig. 311 or Whitehouse, igioa, fig. 5), each of the last several vertebrae articulates with the base of a single caudal ray by means of a hypural (which is probably a hemal spine fused with the radial element of the fin ray, cf., Totton, 1914 : 255). In the teleostean fishes, however, complications of several types occur. In the first place, several of the posteriormost vertebral centra disappear (see below). Second, the hypural elements themselves become reduced in number, presumably through loss and fusion, so that in teleosts there are usually two to several caudal rays articulating with each hypural. Third, one or more elements of the caudal skeleton above the vertebral column lose their proximal contact with the neural arches (this is also true of Amia) becoming the so-called epurals. Finally, in the more primitive isospondylous fishes there are one to four pairs of bones that lie along the posterior end of the spinal cord ; these are the uroneurals. A final term, urostyle, has been used in the literature to cover such a wide variety of structures that it will be completely avoided here. Supposedly the urostyle represents the last upturned vertebral centra. However, Regan (igioa) and Ramanu- jam (1929) have shown that the so-called urostyle of the Clupeidae is mostly formed of uroneurals (see footnote 2 of the present paper), and the uroneurals may form a large part of the " urostyle " of at least some higher teleosts. Another difficulty lies in the fact that at least two of the centra at the end of the vertebral column in such a fish as Flops disappear as separate entities in the higher teleosts. Thus the last visible centrum of one fish is not always the homologue of the last centrum of another. To get around this difficulty it is necessary to establish some reference point that is both homologous and identifiable in as many teleosts as possible. The most satisfactory seems to be the posteriormost vertebra that is consistently present in all isospondylous fishes (except for such aberrant forms as Coilia) and is usually present in other teleosts ; this is here termed the terminal vertebra (TV of Text-figs. 1-15). It can usually be identified by a series of criteria (though the misimpression should not be given that any or all of these will easily or even securely establish the terminal vertebra of all fishes). The best single feature is probably the nature of the anteriormost hypural articulating or fused with it. In most lower teleosts (and indeed in many percoids) the hypurals to the lower caudal lobe are quite constant in number (three) and general shape. The uppermost A CLASSIFICATION OF MODERN I S OSPON D YLOUS FISHES 329 of these (HY3 of Text-figs.) is a rather narrow strut that is never expanded posteriorly. The next hypural below (HY2 of Text-figs.) is broadly wedge-shaped, i.e., consider- ably expanded toward the rear. Below this again is another strut-like hypural with a broad basal articulation or area of fusion with a centrum ; throughout the paper this will be considered the lowermost hypural (HYi of Text-figs.), and the vertebra with which it articulates is the terminal vertebra. This hypural, or hypural I, almost always differs from those behind and above it in having on each side near its base a lateral flange, which usually terminates posterolaterally in a sharp point. (This flange serves for the attachment of part of the flexor caudalis ventralis superficialis muscle (Greene and Greene, 1914 : 46) and probably in part as protection for the blood vessels that exit from the vertebral column behind the base of hypural I. In these respects hypural i is the terminal structure retaining a number of hemal arch characteristics.) Generally, the terminal vertebra may also be identified in other ways. Whatever its relationship with centra (if present) behind it, there is always a normal, hourglass-shaped centrum and a typical intervertebral articulation immediately in front of it. More diagnostic perhaps is that if the front pair of uron- eurals wedge into or fuse with any centrum, it is that of the terminal vertebra (except apparently in osteoglossoids). In some recent teleosts one or two separately ossified centra may remain visible in the adult ; if so these are here termed postterminal 1 centra (PT of Text-figs. 1-4, 6, 7, 10 and 13-15) and are numbered from front to rear. When there are two post- terminal centra, hypural i alone articulates or fuses with the terminal vertebra, and hypurals 2 and 3 arise together on the first postterminal centrum (as in Text-figs. 1-4). Vertebrae just anterior to the terminal vertebra are here called preterminal vertebrae (PR of Text-figs.) and are numbered from back to front. The structure of the upper half of the caudal skeleton seems to be more subject to variation than the lower. This is perhaps to be expected as there is phylogenetic- ally a progressive retraction of the vertebral column from the upper caudal lobe. At any rate, in the isospondylous fishes there is some fluctuation, from specimen to specimen as well as from species to species, in the number of uroneurals and epurals, and from genus to genus in the number of upper hypurals. Two other types of ontogenetic and individual variation should be mentioned. One is the abnormal but frequent doubling of the neural or hemal arches attached to any one centrum, even the terminal centrum (Totton, 1914 : 253). The other is the nature of the association between these arches and their centra. In young fishes the neural and hemal arches are wedged into their centra and only later in ontogeny do they fuse with them ; however, in some forms, e.g., Albula, such fusion never does take place. Finally, it should be noted that dried skeletons, which have been used here, show especially the superficial features of the bones whereas stained and cleared specimens show more of the deeper features. Furthermore, for practical reasons 1 Dr. E. H. Ahlstrom informs me that he prefers to retain the term urostyle for those centra with which the hypurals articulate. If the confusion that has arisen from including the uroneurals in the " urostyle " can be cleared away, there is much to be said for this, especially in view of the term " post- terminal centra " with its self-contradictory adjective that is used in this paper. 33 A CLASSIFICATION OF MODERN ISOSPON D YLOUS FISHES dried skeletons are usually prepared from large adults whereas stained and cleared specimens .are preferably small. Thus an investigation based on dried material is likely to emphasize one set of caudal features, that based on cleared specimens another. COMPARATIVE CAUDAL MORPHOLOGY In the following accounts of individual caudal skeletons, a basal group (Diagram i) is dealt with first. The " lineages " of caudal structure presumably arising from these basal forms are then discussed. Finally the component members of these EPI EP2 UN3 FIG. i. Caudal skeleton of A lepocephalus rostratus , x 5%. Stippling indicates interosseous or infraosseous areas. The arrow indicates the midpoint between the lobes of the caudal fin. EP, epural (these are numbered from top to bottom) ; HS, hemal arch and spine ; HY, hypural (hypurals are numbered from bottom to top ; hypural i is really in large part the hemal arch and spine of the terminal vertebra) ; NA, neural arch (only) ; NS, neural arch and spine ; PR, preterminal vertebra (these are numbered from rear to front) ; PT, postterminal centrum (theseare numbered from front to rear) ; TV, termi- nal vertebra ; and UN, uroneural (these are numbered from front to rear). " lineages " are taken up. Superfamily and subordinal names are those adopted in Section 4 of this paper ; for the moment it need only be noted that Chanos, Kneria, Phractolaemm and Cromeria are here included in the Gonorhynchoidei. Failure to mention certain families is due to lack of available material. Of the caudal skeletons examined the following would seem to be among the most primitive 1 : Elops (Hollister, 1936, fig. 14), Alepocephalus (Text-fig, i), Salvelinus (Text-fig. 2), Esox (Text-fig. 3), and Hiodon (Text-fig. 4). All of these have a number 1 Underlying the use of the terms " unspecialized " and " primitive " throughout this paper is the assumption that in teleostean fishes there has been a progressive fusion and loss of parts in the evolution of the caudal skeleton. (According to this concept the most highly simplified caudal skeletons are often the most " advanced " whereas the more complex are frequently the most " primitive ".) A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 331 of unspecialized 1 features in common : the front end of the anterior uroneural overlaps, but neither wedges deeply into nor fuses with, the terminal vertebra ; hypural i is the only hypural articulating or fusing with the terminal vertebra ; FIG. 2. Caudal skeleton of Salvelinus grayi, x 6|. Bases of some of the upper fin rays are included to show gap for posterior uroneurals. hypurals 2 and 3, and these only, articulate with the first postterminal centrum ; and a portion of a second postterminal centrum is always visible. From the type of caudal skeleton represented in these primitive forms the more specialized caudal structures found among isospondylous fishes have presumably arisen. Indeed it seems possible to trace in the caudal skeleton four main " lineages " (I-IV of Diagram i), as will be done here. As to these, Elops and Alepocephalus Stomiatoidei Mormyriformes (Fig. ?) (Fig. 15) Gonorhynchoidei Osmeroidae (Fig. 8) (Figs. 5. II, 12)) Clupeoidae Argentinoidae (Fies. 6. 7> (Fig. 10) One or no postterminal Osteoglossoidae centra in the adult Anterior uroneural either fused with or wedged Albuloidae j (Figs. 13. 14) Hypurals to the upper half of the caudal fin fused into a single into terminal vertebra / plate-like structure Two postterminal centra, hypurals 2 and 3 articu- / Alepocephaloidae Two postterminal centra, hypurals 2 and 3 articulating with post- lating with postterminal / (Fig- 1) terminal centrum 1 centrum 1 / J> Anterior uroneural neither fused Anterior uroneural neither l^r Dallioidae (!) wit h nor wedged into terminal fused with nor wedged ^~ into terminal vertebra Elopoidae Umbroidae vertebra. Salmonoidae Esocoidae Hiodontoidae '(Fig.2) (Fi^.3) (F-gM, Diagram, i ZOOL. 6, 6. 332 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES are considered to be basal members of a stock including the suborders Elopoidei, Clupeoidei, Gonorhynchoidei, and Stomiatoidei ; Salvelinus represents the Sal- monoidei ; Esox, the Esocoidei ; and Hiodon, the Osteoglossoidei and Mormyriformes. Of these four presumed lineages, the caudal skeletons of adult members of the Salmonoidei (II) are, with the exception of those of the Salangidae and of the neotenic aplochitonid Lovettia, the most easily recognizable. In this suborder the last few preterminal vertebrae have neural and hemal spines with flattened, anteroposteriorly oriented blades (Text-fig. 5). These together tend to make up a flange or keel running -UN V-HY7 FIG. 3. Caudal skeleton of Esox lucius, x 2. above and below the posteriormost portion of the vertebral column. This structure is developed to a variable degree in the different salmonoid fishes but is only totally absent in Lovettia and the salangids ; it is not found in the members of any of the other lineages, though Albula approaches it. In addition there is a series of about 10 small, more or less s-shaped accessory (procurrent) rays in front of the caudal fin above and below in the salmonoids including Lovettia and the salangids ; these salmonoid accessory rays are easier to recognize than they are to differentiate from those that occur in many other isospondylous fishes. The osteoglossoid-mormyriform lineage (IV) has two quite different types of caudal skeleton, the one represented by Hiodon (Text-fig. 4) and the other by Osteo- glossum (Text-fig. 13), Heterotis (Text-fig. 14), and Mormyrus (Text-fig. 15). It would be difficult to confuse the second type with any other isospondylous caudal A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 333 skeleton, but it is quite possible that Hiodon, by a different route, may have produced a fish with a caudal skeleton similar to that in the clupeoid stock. The main caudal feature militating against a hiodontid derivation for such a fish as Gonorhynchus is that this genus and the great majority of the other fishes here assigned to Lineage I have 17 branched caudal rays, whereas all of the members of the osteoglossoid- mormyriform lineage including Hiodon have 16 or fewer branched caudal rays. ^NS of PR .NS of TV -NSofPT PT2 FIG. 4. Caudal skeleton of Hiodon alosoides, X 4 . In this fish the preterminal and terminal vertebrae and the first postterminal centrum each have a neural arch and spine that articulates with its centrum between the lateral uroneurals of the two sides. The haplomous lineage (III), as here understood, contains only the Esocidae, Dalliidae, and Umbridae. Chapman's (1944) assignment of the southern hemisphere aplochitonids, retropinnids, and galaxiids to the haplomous fishes is not accepted partly because of the caudal skeleton but also for other reasons that will be developed in Section 2 of this paper. The caudal skeletons of Esox (Text-fig. 3), Umbra and Novumbra are very similar and primitive. No caudal skeleton of Dallia has been available for investigation. The final lineage recognized here is that including the clupeoids and various other groups (I). About the only feature held in common by all its members is the high number (17 except in Coilia, Phractolaemus , and Cromeria) of branched rays in the caudal. This would seem to distinguish them from the osteoglossoid stock. : 34 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES The absence of antero-posteriorly directed flanges on the last few neural and hemal spines would in turn separate them from the salmonoids. Nevertheless, Lineage I as accepted here may prove to be a heterogeneous assemblage. For example, such groups as the Gonorhynchoidei and Stomiatoidei may very possibly have arrived at caudal skeletons similar to those of the Clupeoidae by a process of parallel or convergent evolution rather than because of genetic relationship : about all that can be said for these questionable groups is that their caudal skeletons do not seem to show relationships with the salmonoid, esocoid, or osteoglossoid lineages. FIG. 5. Caudal skeleton of Hypomesus olidus, x iof . Suborder Elopoidei. Caudal skeletons of Elops, Megalops, Albula and Pterothrissus have been examined. All except the last have been discussed and figured by Hollister (1936, figs. 14-39). Though all are primitive, there is considerable difference between Elops and Megalops on the one hand and Albula and Pterothrissus on the other. In Elops and Megalops there are four pairs of uroneurals overlapping one another rather in the fashion of the body scales ; in the adult Albula and in Pterothrissus there are only two pairs of uroneurals and these are arranged almost linearly, one pair behind the other. In Elops and Megalops the terminal vertebra bears a neural arch and behind it lies a small median crest above and between the uroneurals of each side ; in Albula and Pterothrissus the terminal vertebra bears no neural arch and the median crest behind it has greatly expanded. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 335 All of the species of Elopoidei appear to have 17 branched rays. Elops and Albula, but not Pterothrissus and Megalops, have a single bony fulcral plate at the front of the caudal base above and below. Superfamily Alepocephaloidae. The caudal skeleton of Alepocephalus (Text-fig, i) shows a peculiar combination of primitive (see above) and specialized features. There are the usual 17 branched caudal rays. Only one of the uroneurals is elongated, but this reaches forward to cover part of the first pre terminal vertebra. Parts of two posttenninal centra are visible, and hypurals 2 and 3 articulate with the first of these. Of specializations, the most notable perhaps is the long, curved, rod-like form of the upper epural and of the posterior neural spine. FIG. 6. Caudal skeleton of Chirocentrus dorab, x 4! Superfamily Clupeoidae. The caudal skeleton of Chirocentrus illustrated here (Text-fig. 6) differs from that of the other clupeids and engraulids examined (aside from Coilia) only in the somewhat enlarged and long-based neural arch on the terminal vertebra and the very slight development of a lateral spine on hypural i (cf., Text-fig. 7). A specimen of Nematolosa examined, and illustrations of Clupea given by Ramanujam (1929, fig. 27) and of Anchoviella, Harengula, Opisthonema, and Sardinella by Hollister (1936, figs. 45-53) agree with Chirocentrus in all basic features. In these fishes the anterior uroneural is fused proximally with the terminal vertebra ; hypural 2 has no basal articulation ; one postterminal centrum is visible ; and the neural arch of the terminal vertebra has a dorsally projecting flange that is usually higher than broad. Many of these features have been developed in other isospondylous fishes but the loss of a basal articulation for hypural 2 apparently occurs only in portions of Lineage I. Normally, hypurals 2 and 3 articulate close together on a single centrum ; 336 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES if two postterminal centra are present hypurals 2 and 3 articulate with the first, which would consequently indicate that when two postterminal centra are present the first is of double origin. In the herring, however, Ramanujam (1929) has shown by embryological investigation that it is the terminal centrum that is of double origin (it seems to be formed by the fusion of the terminal vertebra with the front half of the first postterminal centrum), and that presumably with this change hypural 2 has lost its basal attachment. Though the anterior half of postterminal centrum i apparently always fuses into the terminal vertebra in the Clupeoidae, the fate of the more posterior centrum elements seems to vary. Thus, in Clupea, judging from Ramanujam's figures (1929, figs. 26, 27), the posterior half of postterminal centrum i fuses with the base of hypural 3 ; the same thing has probably occurred in Chirocentrus (Text-fig. 6) UN2 -HY3+PT FIG. 7. Terminal vertebra and associated structures of Dussumieria acuta, x and the majority of Clupeoidae. In Chirocentrus, therefore, the element labelled PT is probably postterminal centrum 2, fused with such centrum elements as may occur behind it. From Hollister's figures (1936, figs. 40-45) it would appear that in the dussumieriid Jenkinsiella and perhaps in the engraulid Anchoviella, by contrast, the whole of postterminal centrum i fuses with the terminal vertebra. Finally, in Dussumieria it seems that the posterior half of postterminal centrum i has fused not only with the centrum elements behind it but with the base of hypural 3 as well. Dussumieria also differs from the other Clupeoidae examined in the fusion of hypural i with the terminal vertebra and the anterior uroneural. Superfamilies Chanoidae, Phractolaemoidae, and Knerioidae. The caudal fin of Chanos differs in no very essential way from that of the herring-like fishes. The uroneural is completely fused at its base with the terminal vertebra, but the first hypural is merely suturally articulated. In the single large specimen examined neither hypural 2 nor 3 has any connection with a centrum. The one significant difference from the herrings would appear to be that the median projection between the basal wings of the uroneurals is low, and just above it, extending along the A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 337 upper surface of the uroneural is a single long rod, presumably an epural. This structure is quite reminiscent of that in Alepocephalm (see Text-fig, i) though little else about the caudal skeleton is similar. The caudal skeletons of Phractolaemus and Kneria are in turn very similar to Chanos, differing from that of the herrings in approximately the same way as Chanos does. However, there are only 13 branched rays in the rounded caudal of Phracto- laemus. Super/amity Gonorhynchoidae. The caudal skeleton of Gonorhynchus (Text-fig. 8), as might be expected from the general consolidation of skeletal parts, shows a very FIG. 8. Caudal skeleton of Gonorhynchus greyi, x high degree of fusion for the isospondylous fishes. Not only are the uroneurals ankylosed to one another and to the terminal vertebra along with hypural i, but hypurals 2 and 3 appear to have fused with one another. In general appearance the caudal skeleton of Gonorhynchus bears considerable resemblance to that of Dussumieria, but whether this is a result of convergent evolution or of genetic relationship remains in doubt. In any event, the caudal skeleton of Gonorhynchus gives no indication of relationship with the Osteoglossoidae, and its 17 branched rays would militate against a possible derivation from Hiodon. Suborder Stomiatoidei. Among the stomiatoid fishes the caudal skeletons of Pho- tichthys (Text-fig. 9) and Gonostoma were the only ones available. These two differ considerably from one another. In Photichthys neither the anterior uroneural nor hypural i is fused with the terminal vertebra, but in Gonostoma both appear to be. In both genera there are the usual 17 branched caudal rays. 338 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES Superfamily Scdmonoidae. The salmonids, thymallids, and coregonids (these terms are used here for groups of genera) all have a very characteristic and primitive caudal skeleton (Text-fig. 2). The anterior uroneural on either side bears a high crest between which the front ends of the epurals project. Two postterminal centra are visible ; hypurals 2 and 3 articulate with the first of these, but the base of hypural 2 also extends somewhat forward on to the terminal vertebra. Salmo and Salvelinus have a well-developed neural spine on the first preterminal vertebra (Text-fig. 2), but in two skeletons of Thymallus (see also Lotz, 1864, pi. 10) and one of Coregonus FIG. 9. Terminal vertebra and associated structures of Photichthys argenteus, x 7^. the epurals extend forward over but do not connect with the neural arch of the preterminal vertebra, thus replacing the neural spine in Salmo. Superfamily Argentinoidae. Among the Argentinoidae the only skeleton available is one of Argentina. This is here illustrated (Text-fig. 10) despite the strong possibility that the articulation between hypural I and the terminal vertebra is abnormal. Other available figures of the Argentinoidae are those of Opisthoproctus (Trewavas, I 933. pl- 2) and of Macropinna and Nansenia (Chapman, 19426, fig. 6 and 1948, fig. 12 respectively). All of these show the typical salmonoid neural and hemal spine expansions. They may, however, be separated from members of the Salmonoidae by the fact that the anterior uroneural never extends forward of the terminal vertebra and from the Osmeroidae by the presence of at least one postterminal centrum. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 339 Subsequent examination of two stained specimens, 38 and 46 mm. in standard length, of the bathylagid Leuroglossus stilbius (kindly loaned by D. M. Cohen), shows the same basic structure described above. However, the salmonoid flanges are entirely lacking on the smaller and only slightly developed on the larger example. As compared with Argentina, there are, in Leuroglossus, only two uroneurals, which are very slender, and the anterior of these is fused with the terminal vertebra. The postterminal centrum is considerably smaller than that of Argentina (Text-fig. 10), and only hypural 5 articulates with it. Also, Leuroglossus has no epurals. EPI EP2 UN3 HY7 HY6 FIG. 10. Caudal skeleton of Argentina silus, x Super family Osmeroidae. Of the members of the Osmeroidae mentioned below, all but the Salangidae and Lovettia have the typical salmonoid neural and hemal spines (Text-fig. 5). None of them show a postterminal centrum or for that matter any other caudal resemblance to the esocoids. Though there is considerable variation among the members the most striking skeletal difference is that in Retropinna and Galaxias (Text-fig, n) the anterior uroneural is not fused with the terminal vertebra whereas in Hypomesus (Text-fig. 5), Plecoglossus (Text-fig. 12), Aplochiton, and the salangid Leucosoma it is. The caudal skeleton of the last-named genus is quite different from all of the others. In one respect the caudal fins of the southern Osmeroidae differ from those of the northern members of the group (also from the Argentinoidae and Salmonoidae) . All of the latter, including the salangids and Plecoglossus, have 17 branched caudal ZOOL. 6, 6. 2I 340 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES FIG. ii. Caudal skeleton of Galaxias fasciatus, x 6|. HY7 FIG. 12. Caudal skeleton of Plecoglossus altivelis, x 8. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 341 rays. In the southern Retropinna and Prototroctes there are 16 branched caudal rays ; and in Aplochiton, Lovettia, and Galaxias 14. Suborder Esocoidei. The caudal skeletons of Esox (Text-fig. 3) and of Umbra are quite similar. That of Umbra appears the more primitive of the two, with six " hypurals " of about equal size between a narrower strut above and below. The upper of these struts is the uppermost hypural ; the lowermost is the hemal arch from the second preterminal vertebra, the first preterminal vertebra giving rise to one of the broad " hypurals ". In seven specimens of Esox lucius for which caudal rays were counted, five had 17 branched rays and two 16 ; for three specimens of Umbra krameri, two had 9 branched rays and one 8. FIG. 13. Caudal skeleton of Osteoglossum bicirrhosum, x Superfamily Hiodontoidae, As previously mentioned, one of the most primitive of isospondylous caudal skeletons is certainly that of Hiodon. In two skeletons examined, one of H. tergisus and one of H. alosoides, considerable variation appeared, not only from fish to fish, but on the two sides of the same fish. Thus the skeleton of H . alosoides has three uroneurals on one side (Text-fig. 4) and four on the other. In H. tergisus both the terminal and the first postterminal centra have double neural arches, and the anterior tip of the single epural runs in between but is free from the sides of the posteriormost arch. In H. alosoides (Text-fig. 4) both the terminal and the first postterminal vertebrae have single neural arches with spines that nearly reach the caudal fin base. Nine specimens of H. alosoides and one of H. tergisus examined all had 16 branched caudal rays. Super families Osteoglossoidae and Notopteroidae. The caudal skeletons of Osteo- glossum (Text-fig. 13), Pantodon, and Heterotis (Text-fig. 14), while showing strong 342 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES similarities to one another, are so different from those of all the foregoing isospondy- lous fishes as to be somewhat difficult to interpret. (Notoptems, see Whitehouse, 19106, pi. 47, fig. 6 is even more specialized in the same direction, and a skeleton of Arapaima was not available.) Presumably, the terminal vertebra is that preceding FIG. 14. Part of caudal skeleton of Heterotis niloticus, x 4. FIG. 15. Caudal skeleton of Mormyrus caballus, x the one to which two hypurals are attached. If this is correct, postterminal centrum i is present in an unusually complete form, and postterminal centrum 2 has fused with the large upper hypural plate and with the uroneurals as well (the latter fusion is incomplete in Heterotis, Text-fig. 14). Though any other interpretation of these caudal skeletons raises even more difficulties, the one just given presents three : first, hypural 3 in Osteoglossum has lateral flanges that are not found elsewhere in A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 343 isospondylous fishes ; second, postterminal centrum I in Osteoglossum has a neural spine instead of the usual epural structure ; finally, and most oddly, the uroneurals in all three genera under consideration fuse with postterminal centrum 2 instead of with the terminal vertebra. Because of this last peculiarity, the alternate inter- pretation that what is here called postterminal centrum 2 is really the terminal vertebra deserves serious consideration. In any event, it is difficult to see how the caudal structure of these fishes has been derived from any of the isospondylous groups, including Hiodon, dealt with up to here. Heterotis has only 12 branched caudal rays, and Osteoglossum 14. Order Mormyriformes. The caudal skeleton of Mormyrus caballus (Text-fig. 15) is basically very similar to those of Heterotis and Osteoglossum. The main differences lie in the flange on hypural 2 and the number of caudal rays in Mormyrus (16 branched). DISCUSSION Two subjects call for discussion. One is the functional significance of the different types of caudal skeletons noted above. The other concerns the bearing of the caudal structure on the limits of the order Clupeiformes. Two aspects of the relation between structure and function will be considered 1 . The first of these concerns the elongate uroneurals of the primitive isospondylous fishes. According to Regan (igioa) and Ramanujam (1929) the uroneurals are, at least primarily, modified portions of neural arches of postterminal centra. What- ever their derivation, the elongation of these uroneurals in the so-called primitive isospondylous caudal skeleton is in itself a specialization. The only obvious explana- tion for these long uroneurals is that they stiffen the upturned terminal portion of the vertebral column. With the more abrupt upturning in this area found in advanced members of both the clupeoid and salmonoid lineages, the uroneurals, which start out as one or more shields along the sides of the postterminal centra, develop into struts which are at first wedged into the terminal vertebra and in more " advanced " forms fuse with it. Apparently a short, anchored strut serves here better than a long cover. The second functional aspects has to do with the relationship between the caudal skeleton and the shape of the tail. In a very broad sense there is a correlation between these two features. Very generally, when the tail is forked the caudal skeleton is divided into distinct upper and lower portions, but when the tail is rounded the posterior outline of the hypurals is rounded, often with a number of hypurals of subequal size (e.g., Amia, Umbra] or with a central one somewhat larger than the others as in the cods. However, this relationship is far from strict. For example, the round-tailed serranid Epinephelus has the caudal skeleton of a typical fork- 1 In this discussion the caudal skeleton is treated as if it were an independent functional unit. Ob- viously it is not, for beside providing an axis for the basal articulation of the caudal rays it serves as a source of attachment for some of the caudal musculature (cf. Hindersson, 1910 ; Schmalhausen, especially 1913 ; and Greene and Greene, 1914) and for the transmission of the blood vessels and nerves to the caudal fin. However the role that these various functions have played in the evolution of the caudal skeleton are almost entirely unknown at the present time and have consequently been left out of con- sideration. 344 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES tailed fish. Conversely, the fork-tailed Mormyrus has a caudal skeleton that should belong to a round-tailed fish. Apparently the shape of the tail takes a very long time to imprint itself on the structure of the caudal skeleton. Similarly there seems to be a relationship, though again not a strict one, between the shape of the caudal fin and the number of caudal rays. In fork-tailed fishes the number is usually constant within groups. Indeed the basic number of 17 branched caudal rays runs through two of the four isospondylous lineages postulated here. However, a rounded caudal is often associated with a lower (and variable) number of principal caudal rays. Presumably, in a fork-tailed fish the greatest strength and functional importance lies in the outermost principal caudal rays which form rigid limits to the fin. In round-tailed fishes these outer rays lose their significance and the fin tapers from the middle to above and below. One is inclined to wonder whether those fork-tailed isospondylous fishes with some lower caudal count than the typical one, e.g., Hiodon, Aplochiton, have not had some round-tailed ancestors? With regard to the information caudal structure can provide on the limits of the order Clupeiformes, there are only three groups that need be considered here : the haplomous, mormyriform, and iniomous fishes. Each of these has been included in or excluded from the Clupeiformes, depending on the classification adopted. The first two of these groups have already been dealt with. Suffice it to say here that so far as the caudal skeleton is concerned the haplomous fishes appear to be as primitive as any of the isospondylous fishes. Since they stand at the base it is impos- sible to derive them from any other modern isospondylous fish, so far as the caudal skeleton is concerned. The mormyroid caudal skeleton seems clearly to have been developed from the type now found in Heterotis. The structural inference, as already noted, is that the mormyrids retain not only the imprint established in a round-tailed form, but one of a peculiar and definitely identifiable type. Among the iniomous fishes (Scopeliform.es) only skeletons of Neoscopelus, Lam- panyctus, Aulopus, and synodontid fishes have been available for examination. The caudal structure of these seems to be somewhat more primitive than that of many isospondylous fishes in that one postterminal centrum is clearly visible. In certain features, e.g., the flanged uroneural, these iniomous caudal skeletons appear to approach the salmonoid Clupeiformes ; however, they lack the expanded posterior neural and haemal spines of the salmonoids. In sum, there is little beyond an indica- tion of possible relationships here. Whether any or all of the three groups discussed above should be included in or excluded from the order Clupeiformes will, of course, depend in the first place on the nature of the whole animals, not just their tails, and in the second on a consensus of human opinion regarding what constitutes a fish order. SECTION 2. THE CLASSIFICATION AND RELATIONSHIPS OF THE SALMONOID FISHES The members of the Suborder Salmonoidei (as defined here) today are the dominant fishes in the coldest fresh water of both hemispheres. They are also represented in A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 345 the temperate and arctic regions of northern seas by one group (Osmeridae) and in deeper oceanic waters by another (Argentinoidae 1 ). It is hardly surprising that there should be difficulty in denning as large and varied an assemblage as the salmonoids. Regan (1929) and others have differentiated them from the clupeoid groups by the absence or incompleteness of the oviducts. However, Kendall (1921) demonstrated that the oviducts of Salmo and Osmerus " are not radically different from those of other isospondylous fishes ". Regan (1913^ : 289) gives the following additional salmonoid characters : an adipose fin usually present, one supramaxillary or none, and parietals well developed. Probably the most useful of these three features for purposes of differentiation is the adipose fin, for it is found elsewhere in the isospondylous fishes only among the stomiatoids and the Rosauridae. However, there are salmonoids (Galaxias) and salmonoid derivatives (the haplomous fishes) in which the dorsal fin has moved far back on the body and presumably squeezed out the adipose fin (Garstang, 1931 : 246-250). An adipose is also lacking in at least one of the oceanic salmonoids (Microstoma) . An additional diagnostic trait of most salmonoids is that the last few " preterminal " vertebrae have laterally compressed neural and hemal arches, forming a flange above and below the posterior portion of the vertebral column (see previous section). This flange seems to be lacking only in the neotenic Lovettia, and the Salangidae among the salmonoid skeletons examined. At least the posteriormost of these neural and hemal spines lead out to a series (usually about 10) of characteristic small, curved, accessory rays ; these occur prominently in both Lovettia and the salangids. A final trait that will apparently distinguish the salmonoids (and haplomous fishes) from the herrings and their relatives is the absence of a lower series of intermuscular bones, i.e., the series that articulates proximally with the base of the lower ribs (cf., Emelianov, 1935.) The present classification of the salmonoid families stems from Regan (19130 : 289). There, he differentiates the families as follows : "I. An orbitosphenoid ; an opisthotic ; a mesocoracoid ; vertebrae upturned at base of caudal fin i. SALMONIDAE II. An orbitosphenoid; no opisthotic; no upturned vertebrae; mesopterygoids toothless. A mesocoracoid ; parapophyses inferior . . . . . 2. ARGENTINIDAE No mesocoracoid ; parapophyses lateral .... 3. MICROSTOMIDAE III. No orbitosphenoid ; no opisthotic ; no upturned vertebrae ; mesopterygoid toothed (absent in the Salangidae). A. A mesocoracoid ; maxillaries dentigerous, entering gape . . 4. OSMERIDAE B. No mesocoracoid ; maxillaries dentigerous, entering gape. Head compressed ; mesopterygoid well-developed dentigerous ; ribs ossified 5. RETROPINNATIDAE Head strongly depressed ; no mesopterygoid ; ribs not ossified . 6. SALANGIDAE C. No mesocoracoid ; maxillaries toothless, behind praemaxillaries. Praemaxillaries not extending whole length of maxillaries ; roof of myodome unossified ; no adipose fin. ....... 7. GALAXIIDAE Praemaxillaries nearly reaching extremities of maxillaries ; roof of myodome ossified ; an adipose fin . . . . . .8. HAPLOCHITONIDAE" 1 The name Argentinoidae, rather than Opisthoproctoidae, will be used throughout this paper for the perhaps over-simple reason that the most recent workers on the group (Hubbs, 1955 ; Bertelsen, 1958 ; and Cohen, 1958) seem to prefer the former name. 346 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES With regard to this classification, Regan goes on to say (op. cit.: 290) : "It is of some interest to note that the Galaxiidae and Haplochitonidae are related to, but more specialized than, the Osmeridae, or Smelt family, of northern seas. Retropinna, from the coasts and rivers of Australia and New Zealand, is still nearer to the Galaxiidae and Haplochitonidae." Subsequent work by Trewavas (1933), Parr (1937), and Chapman (19420, 19426, 1943, and 1948) has abundantly shown that the members of Regan's section II including his Argentinidae and Microstomidae are rather widely divergent from the other salmonoid fishes. Characteristic features of the group are the toothless upper jaw, the pelagic eggs, etc. The classification of the families within the group seems to be rather controversial (cf., the above papers, Hubbs, 1953, and Bertelsen, 1958) but the subject will not be entered upon here. Similarly the members of Regan's (19130) section I may be readily distinguished from all other members of the suborder by the three upturned caudal vertebrae. (Other members of the suborder have one or, at most, two.) Whether the family Salmonidae should be interpreted in the broad sense of Regan (19130) or whether the Thymallidae and /or Coregonidae should be split off as separate families are matters that can best be left to those more familiar with these fishes to decide. However, the assignment of the fossil Thaumaturidae by Voigt (1934), Berg (1940) and others to the salmonid group seems to be open to some question. The typical salmonoid flanges formed by the last few neural and hemal spines are lacking, as is the adipose fin (ace. Voigt) ; the upward inflexion of the last three vertebrae shown in Voigt's pi. 2, fig. 5 is not salmon-like and is, in any event, not duplicated in the specimens shown in pi. i, figs. I and 2 ; and the posterior position of the dorsal is more reminiscent of the haplomous fishes than of the salmons. The last group of salmonoid fishes section III of Regan contains a rather diverse assemblage of fishes. It can, however, be immediately divided into a northern group of families Osmeridae, Plecoglossidae, and Salangidae with 17 branched caudal rays, and a southern hemisphere group Retropinnidae, Aplochitonidae, and Galaxiidae with fewer branched caudal rays. This subdivision is in line with Regan's (19130) statements quoted earlier. Nevertheless both groups are quite varied within themselves, and each has given rise to a series of specializations partly paralleled within the other. Thus the northern Plecoglossus with its specialized dentition to some extent resembles the southern Prototroctes with its small flexible teeth, and the northern neotenic salangids are partly duplicated in the southern neotenic Lovettia. The salmonoid allocation of the three southern hemisphere families remained uncontested until Chapman (1944), on the basis of an osteological study of Aplochiton, transferred the Aplochitonidae, Galaxiidae, " and very probably the Retropinnidae " to the haplomous fishes. Subsequent authors, e.g., Schultz and Stern (1948) and Wilimovsky (1951), have generally accepted Chapman's allocation. However, Blackburn (1950), from an investigation of the biology of the aplochitonid Lovettia seali, points out " that the life history of this species is similar to that of various salmonids, which makes the author incline to the views of Regan and Jordan " with regard to its systematic position. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 347 In view of this difference of opinion, the British Museum material of the southern hemisphere forms has been examined. A general review of these, at least insofar as their systematic position is concerned, is undertaken in the following paragraphs. In general the three southern families can be divided into two groups the anadro- mous retropinnids and aplochitonids (Prototroctes?) with an adipose fin and a forked caudal, and the fresh-water or catadromous galaxiids without an adipose fin and with a rounded caudal. The galaxiids would appear to be a fairly homogeneous group even though they are represented in southern South America, southern Africa, New Zealand, Tasmania, Australia, and presumably New Caledonia. Regan (1905) has an early revision of the family ; more recently Scott (1936, etc.) and Stokell (1945, etc.) have worked on the systematics of the group. Swinnerton (1903) has written on the osteology of one of the species. By contrast with the galaxiids, the retropinnids and aplochitonids are most diverse. The Retropinnidae has only a single Australian and New Zealand genus revised by Stokell (1941). The Aplochitonidae contains three very different genera : a fairly normal Aplochiton from southern South America and the Falkland Islands, Proctotroctes with a highly specialized dentition and with two almost if not entirely extinct species from Australia, Tasmania, and New Zealand, and the neotenic Lovettia from Tasmania. As already noted, Chapman (1944) has a paper on the osteology of Aplochiton, and Blackburn (1950) has published on the biology of Lovettia. A general account of the New Zealand forms of all three families is to be found in Stokell (1955). In external characters, aside from those already mentioned, Retropinna and Proctotroctes are scaled ; the other southern genera are naked. Retropinna has somewhat the look of, and is called, smelt (Stokell, 1955 : 9-18, pis. 4 and 5) whereas Prototroctes looks rather like a Coregonus (Stokell, 1955 : 41-44, pi. 13). Aplochiton is more trout-like (Eigenmann, 1927 : pi. 15) ; Lovettia has a definitely larval appearance (Blackburn, 1950, pis. i and 2) ; and the galaxiids (Stokell, 1955 : 19-40, pis. 6-12) have somewhat the look of the northern Umbra. So far as I can see the rather variable fin counts have no significance for the classi- fication of the group as a whole except for the caudal. Judging from counts in a single specimen of each genus, Prototroctes and Retropinna have 16 branched rays, whereas Aplochiton, Lovettia and Galaxias have 14 branched. All of these genera have the typical salmonoid accessory rays (see previous Section) at the front of the caudal above and below. The dentition of these southern hemisphere forms is most varied. At the one extreme are certain species of Retropinna (Stokell, 1941) in which there is a long posterior portion of the maxillary, a short premaxillary, a crescent-shaped vomer, palatines, mesopterygoid, tongue, a long bone covering the basibranchials, pharyn- geals, and dentary all with large teeth. At the other extreme is Prototroctes with fine, peculiarly-shaped, flexible teeth in the upper jaw and apparently none in the lower. The maxillary is also variable. In the single genus Retropinna it may be toothed or toothless (Stokell, 1941 : pi. 55). In the three genera of the family Aplochitonidae 348 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES the toothless maxillary is excluded from the gape as effectually as it is in the perch- like fishes. A supramaxillary seems to be consistently absent. In Galaxias and Lovettia the premaxillaries appear to be slightly protractile ; in both these genera they have rudimentary anterior pedicels. With regard to the skull the parietals meet on the midline except in Lovettia. In a ripe female specimen of Lovettia both the frontals and parietals are elongate bones running along the superolateral borders of the skull. The whole median area of the skull roof is merely membranous, with the brain clearly visible. The anterior ends of the frontals project well forward over the ethmoid, not only in Lovettia but in all the other southern genera as well. There are no proethmoids in any of them and what Chapman describes as proethmoids in Aplochiton (1944 : 150, fig. i) are not separate elements in British Museum specimens but are merely anterior prongs of the frontals. (If these anterior portions somehow came to develop as separate elements in the northern haplomous fishes, would they not be proeth- moids?) Whether the mesethmoid is cartilaginous or bony in Lovettia is impossible to say ; it is certainly an endochondral formation. However, in Galaxias (cf., Swinner- ton, 1903) there is a dermal mesethmoid as well. I have not seen an orbitosphenoid in any of the southern hemisphere forms. At the back of the skull the small supraoccipital is excluded from the foramen magnum in all the southern genera. In the suspensorium the large mesopterygoid is usually toothed in southern forms. The metapterygoid extends forward over much but not all of the quadrate. There seems to be no knob behind the palatine for articulation with the lateral ethmoid (prefrontal) . As to the ectopterygoid, it and the palatine seem to be fused into a single long, mostly toothed bone in Retropinna. However in Prototroctes the two elements are separate. The suboperculum is a rather large bone forming part of the gill cover in all except Lovettia. In Lovettia, the suboperculum gives the distinct impression of being the seventh (and terminal) branchiostegal. In the pectoral girdle, none of the southern forms has a mesocoracoid. In the caudal skeleton there are never any visible " postterminal centra " (see previous section). The anterior uroneural may (Galaxias) or may not (Prototroctes) be fused with the "terminal vertebra". In all, except Lovettia, the salmonoid flattening of the last few neural and hemal arches is distinctly developed. As to soft anatomy, oviducts appear to be lacking in the females. In adult males of Lovettia, which seems to be unique in this respect, the urinogenital openings move forward to just behind the pectorals. None of the southern genera have pyloric caeca. Lovettia has a closed air bladder quite unlike that of the osmerids, salmo- nids, or northern haplomous fishes including salangids (N. B. Marshall, personal communication). With this background regarding the southern hemisphere Aplochitonidae, Retro- pinnidae, and Galaxiidae, it remains to compare them with the northern hemisphere Esocoidei and Osmeroidae. The northern haplomous fishes have been defined by Chapman (1934 : 372) as A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 349 follows (and the definition has been accepted with little change by Berg, 1936, 1940, and 1955) : " Physostomous teleosts with the pelvic fins abdominal ; cycloid scales ; paired proethmoids ; parietals separated by the supraoccipital ; maxillary forming part of the edge of the upper jaw but lacking teeth ; premaxillary not protractile ; head scaly ; the dorsal and anal fins far back on the body ; but without adipose fin, mesocoracoid, or orbitosphenoid, and with no true spines in the fins." As has already been noted, the southern hemisphere Lovettia has a closed air bladder ; none of the southern forms have paired proethmoids ; only Lovettia has the parietals in any degree separated by the supraoccipital ; the maxillary is excluded from the gape in the Aplochitonidae, and in some species of the Retro- pinnidae the maxillary is toothed ; in Lovettia and Galaxias the premaxillary appears to be slightly protractile ; none of the southern forms have scales on the head ; the dorsal in Proctotroctes is not especially far back ; and only the galaxiids lack an adipose. If, then, the southern hemisphere families are included in the Esocoidei, only the following portions of Chapman's (1934) definition are left intact : " Teleosts with the pelvic fins abdominal ; cycloid scales ; anal fin far back on the body ; without mesocoracoid or orbitosphenoid, and with no true spines in the fins." The above words would apply, except for the absence of a mesocoracoid, to most Osmeroidae, and in toto to the included family Salangidae. This is not to say that the southern families do not have haplomous resemblances, for they do. Indeed, some, but by no means all, of the similarities between Aplochiton and the haplomous fishes which Chapman lists (1944 : 164) will hold for all three southern families. Two more may be added here. First the southern families, like the northern Esocoidei, lack pyloric caeca, whereas the northern osmeroids (except the Salangidae) have several to very numerous caeca. Second, the southern forms all have fewer than 17 branched caudal rays, whereas the northern osmeroids all have 17 branched rays ; in the northern haplomous fishes there is a reduction in caudal ray number except in Esox, and even here two of the seven specimens examined had fewer than 17 branched rays. In comparing the southern families with the northern Osmeroidae, Chapman (1944 : 163) has already noted a number of differences, and several more may be added. The most important would seem to be the following : in the north a supra- maxillary, mesocoracoid, epipleural ribs, and pyloric caeca are present (except in the Salangidae), whereas in the southern forms they are absent ; and in the northern forms there are 17 branched caudal rays whereas there are 16 or fewer in the southern families. Admittedly these differences in the southern group are all modifications away from the basic salmonid type and in a haplomous direction. The same may be said of the posterior dorsal and anal position and the loss of an adipose in the galaxiids, which, of all the southern forms, have progressed farthest along this route. However, this direction has also been followed to a considerable extent by the northern salangids. If, however, it is admitted that all three southern families are related and if the 350 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES most primitive features of these families are added together, we have a fish that is basically osmerid. In Prototroctes the dorsal is median and the adipose is well developed. In some species of Retropinna the maxillary is toothed and indeed the whole dentition is osmerid, with the teeth of the tongue and basibranchials biting between those of the palatines and mesopterygoids. At the other end of the fish the salmonoid flanges on the neural and hemal spines are present (except in Lovettia) OSMERIDAE and PLECOGLOSSIDAE proethmoids sometimes present SALANGIDAE no pyloric caeca no mesocoracoid no flanges on neural and hemal spines APLOCHITONIDAE RETROPINNIDAE and GALAXIIDAE 16 or fewer branched caudal rays no pyloric caeca no mesocoracoid adipose fin present or absent maxillary toothed or toothless two postterminal caudal centra SALMONOIDAE 17 branched caudal rays no proethmoids pyloric caeca present mesocoracoid present an adipose fin a toothed maxillary flanges from posterior neural and hemal spines ESOCOIDEI usually fewer than 17 branched caudal rays proethmoids present no pyloric caeca no mesocoracoid no adipose fin no toothed maxillary no flanges from posterior neural and hemal spines Diagram 2 along with the curved accessory caudal rays, and the caudal supporting skeleton is osmerid (see Section i). This last feature seems particularly important in view of the fact that the caudal skeleton of the haplomous fishes is not only very different but far more primitive than that of the northern and southern Osmeroidae. Finally, the anadromous life history of several of the southern forms recalls that of many salmonids. From all this it seems to the present author that the Aplochitonidae, Retropinnidae, and Gal- axiidae are derivatives of a proto-osmerid stock that have evolved, some more and some less, in a haplomous direction ; and that this same direction has been followed a third time to a lesser extent by the Salangidae (see accompanying diagram). A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 351 That the Salangidae and the southern hemisphere Osmeroidae have departed in the same direction from the northern Osmeroidae is indicated by the characters listed in the diagram. But the similarities between the salangids and the aplochitonid Lovettia are even more striking. Both Lovettia and the salangids are more or less neotenic, anadromous fishes, with tremendous sexual differentiation. The sexual differentiation, however, is of a very different type in the two groups : in the salangid males, but not the females, there is a series of large scales above the anal fin, and the fin itself has a peculiar curvature in several of the median rays ; in the adult male Lovettia the urinogenital papilla is just behind the pectoral fins. In the head of both, the posterior portion of the skull is very incompletely roofed by the frontals and parietals. Furthermore, these are the only two groups in the Salmonoidei that have no flanges on the preterminal neural and hemal spines. In view of this it may appear that Lovettia and the salangids are closely related. Again I would prefer the hypothesis of parallel (neotenic) development. That Lovettia belongs with the southern osmeroids, and the salangids with the northern forms is indicated by the caudal ray count. In dentition, the salangids have a toothed maxillary typical of northern osmeroids, whereas Lovettia has a toothless maxillary excluded from the gape typical of the other two genera of aplochitonids. Additional characters for differentiating at least Salangichthys microdon from Lovettia sealei are the presence of a high (cartilaginous?) lump just behind the palatine for articula- tion with the lateral ethmoid (prefrontal) in Salangichthys and the absence of such a structure in Lovettia ; the presence of a well-developed subopercle in Salangichthys whereas the subopercle resembles a free branchiostegal in Lovettia ; the six branchio- stegals of Lovettia vs. the four in Salangichthys ; and finally the normal, open sal- monoid air bladder in Salangichthys vs. the peculiar closed one in Lovettia (see above). A final question regards the interrelationships of the members of the southern group. Berg (1940) separated the galaxiids as a distinct order, primarily on the basis of brain structure. Unfortunately, all that is known about the brain of these southern forms seems to be a single remark by Swinnerton (1903) regarding Galaxias. However, there are other features that set Galaxias apart from the other southern osmeroids. Among these may be mentioned the general body form and the catadromous habits. Another line of differentiation within the southern section of the Osmeroidae is shown by Lovettia with its neoteny, sexual differentiation, and closed swimbladder. Such peculiarities would certainly warrant family differentiation from the aplochi- tonids, except that the Aplochitonidae is already a family of only three genera. Furthermore, the two remaining genera Prototroctes and Aplochiton are also very different from one another, and it seems preferable not to erect three separate families each with a single genus. Nevertheless, the fact remains that the southern Osmeroidae are now represented by several highly differentiated forms. Indeed, the differentiation is as great as in the northern Osmeroidae. So far as the haplomous fishes are concerned, Berg stated in 1936 that the " Eso- coidei are a specialized group of Clupeiformes originating from Osmeroid fishes at the end of the Cretaceous ". The present paper would tend, like that on Aplochiton by Chapman (1944), to demonstrate an even closer set of morphological links between 352 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES osmeroids and esocoids than was known to Berg. However, the primitive caudal skeleton of the haplomous fishes indicates (see previous section) that Esox, Umbra, etc. have been derived from a somewhat less advanced salmonoid stock than the northern and southern Osmeroidae. From the foregoing account, a classification (to superfamily) of the modern salmonoid fishes and their haplomous derivatives may be written as follows : Suborder Salmonoidei Superfamily Salmonoidae Superfamily Argentinoidae (= Opisthoproctoidae) Superfamily Osmeroidae Suborder Esocoidei Superfamily Dallioidae Suprfamily Umbroidae Superfamily Esocoidae In this classification the three haplomous superfamilies of Chapman (1934) and Berg (1940, 1955) have been accepted without any attempt at critical evaluation. As to the salmonoid superfamilies listed here, everyone admits that they are related ; the only question is how to express the relationships. Here, the salmonid and osmerid groups have been recognized as separate superfamilies, and the argentinoids have been demoted to a superfamily equivalent to them. The first action is taken because it seems advisable, where feasible, to segregate a basal stock that has given rise to a number of lineages from the advanced members of those lineages. The place that the Argentinoidae should hold in relation to the other two salmonoid superfamilies is somewhat less clear. Nevertheless it can only obscure the relation- ship that undoubtedly does occur to set the Argentinoidae up as a suborder (Opistho- proctoidei of Berg, 1940, Chapman, 1942^, etc.) with a taxonomic rank equivalent to that of the salmonoids themselves. In the specializations of the head skeleton (Chapman, 1942^ and b), in the loss of a ventral mesentery (Kendall, 1935, Table I), and in the caudal skeleton (see previous section), the Argentinoidae is more speci- alized than the Salmonidae 1 . It would seem that the Argentinoidae is a much more divergent offshoot of the stock represented by Salmo today than is the Osmeroidae. But to express this greater divergence in a classification seems to lead to more incongruities than to consider them as a superfamily equivalent to the Salmonoidae and Osmeroidae. SECTION 3. THE GONORH YNCHOID FISHES AND THEIR RELATIONSHIPS The single modern genus generally attributed to the suborder Gonorhynchoidei is Gonorhynchus (cf., Regan, 1929 ; Berg, 1940 ; Wilimovsky, 1951 ; etc.). The systematic position of the suborder has always been dubious. Ride wood (19056), on the basis of a study of the skull of Gonorhynchus greyi, while emphasizing its isolated position, tentatively suggests a distant relationship with the salmonids. 1 Another distinguishing character usually attributed to the argentinoids is the presence of a spiral valve in the intestine (cf., Kendall & Crawford, 1922). However, Cohen (1958 : 97, 98) has recently demonstrated that the " spiral valve " of argentinoids is not so diagnostic of that group as has usually been stated. A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES 353 Regan (1929), in his synopsis of the isospondylous fishes, places the gonorhychids at the end of the order, behind the mormyrids. His basis for this placement is that the gonorhynchids have the " parapophyses ankylosed with centra, appearing as strong processes". Subsequent authors, e.g., Berg (1940), Wilimovsky (1951), have generally accepted Regan's allocation. That the parapophyses in Gonorhynchus are strong processes fused with the centra is true, but the value of this character in the classification of the isospondylous fishes seems dubious. For example, among the groups included by Regan in the section with Gonorhynchus, this genus and Osteoglossum have " strong processes " but Hiodon and Mormyrus appear to have none whatever, the ribs articulating directly with the centra. A search for other characters that might indicate the relationships of Gonorhynchus discloses the following, all of which point toward an albulid-clupeid relationship. There are 17 branched caudal rays (this could equally well indicate a salmonoid ancestry). The caudal skeleton, except for the greater fusion of parts, resembles that of Dussumieria, the upright neural arch on the terminal vertebra being especially reminiscent of the clupeoids. Arising from the parapophyses are two sets of seg- mental bones ; one of these undoubtedly represents the lower ribs, but the other seems to be the lower intermuscular bones. Lower intermuscular bones again are characteristic of the clupeoids and seem to be absent from most, if not all, other isospondylous groups. As to relationships within the great clupeoid stock, it is here suggested that the gonorhynchids belong with the chanids, phractolaemids, kneriids, and cromeriids. These five groups are so widely different that any relationship between them is difficult to comprehend. Yet the following similarities may be marshalled. In external characters, the gill membranes are always attached to the isthmus, and in the three African families the gill openings are considerably restricted. In all, the mouth is small and toothless, or nearly so. The preopercular border is free only below (Chanos), if at all. In the head the maxillaries run forward behind the premaxillaries and nearly or quite meet on the midline. There are no supramaxillaries. A preorbital bone extends forward above the upper jaw nearly to the end of the snout (this is least developed in Chanos). The preopercle extends only a short way up the side of the head and there may be a suprapreopercular (Phractolaemus and Chanos} or a long tube-containing ossicle (Kneria} above it. The lower preopercular limb, however, runs well forward and forms the chief support for the quadrate. On the skull roof the frontals are large and the parietals are very small and almost restricted to the rear border of the skull, with the supraoccipital separating those of the two sides. The supratemporal commissure of the lateral line either runs across the parietals and supraoccipital through a series of superficial ossicles or through these bones themselves. In Gonorhynchus and Chanos the exoccipitals send out a pair of prongs that roof the anterior portion of the spinal cord ; in the flat-headed Kneria and Phractolaemus the projection of the supraoccipital serves the same purpose, with the walls of the first few neural arches forming the sides. 354 A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES The lower intermuscular bones are present in Chanos and Gonorhynchus, apparently absent in Phractolaemus and Kneria. In Gonorhynchus the lower ribs are short and slight, and articulate with lateral expansions of the centra. In Chanos the ribs are large and their upper portions articulate directly with the centra, the lower by way of a bony nodule. In Phractolaemus and Kneria the ribs articulate directly with the centra but in the former the ribs are very strong, in the latter moderate. Admittedly, these vertebral characters show little relationship between the genera, but they are given because of their previous use in classification. The caudal skeleton of all these groups (Cromeridi} is essentially the same. As to mode of life, Gonorhynchus is today a purely marine fish, but the group to which it belongs is known from Eocene fresh-water deposits from America and Europe. Chanos is euryhaline, and the other three genera are restricted to fresh water. All five genera are tropical. A good many of the similarities in these five groups may be primarily or secondarily related to the small size of the mouth in all these forms. It could therefore be that the small mouth has developed independently two or more times and that all of the other changes have followed as a matter of course. Some check on this possibility is offered by the members of the Argentinoidae, a small-mouthed group in the salmonoid stock. To some extent a parallel evolution is indicated between the Argentinoidae and the five groups taken up here, particularly with regard to jaw structure, suspensorium and the reduction in branchiostegal rays. However, there are a number of features in the five groups that are not duplicated in the Argentinoidae. Among these are the attachment of the gill covers to the isthmus, the loss of the normal upper portion of the preoperculum, the passage of the supratemporal commis- sure over or through the supraoccipital, the expansion of the anterior portion of the spinal canal, and the advanced form of the caudal skeleton. To summarize, the argument presented here is not that Gonorhynchus, Chanos, Cromeria, Kneria, and Phractolaemus are closely related. Indeed, no two of them are. It merely seems to the author that each of the five is more closely related to the others than it is to anything else. If this is granted, then it becomes fairly easy to tie in the group, via Chanos and to some extent Gonorhynchus, with the general clupeoid stock 1 . SECTION 4. A CLASSIFICATION OF THE MODERN ISOSPONDYLOUS FISHES The systematics of the living isospondylous fishes have received a great deal of attention. Because of the excellent literature review by Wilimovsky (1951) only a sampling of papers on the subject need be mentioned here. With regard to the general classification of the groups of Clupeiformes the deserved- ly classical papers are those of Ridewood (19040 and b, 1905^, b, and c). Unfortu- 1 The British Museum (Natural History) material upon which the above account has been based consists of complete skeletons of Gonorhynchus and Chanos, somewhat incomplete skeletons of Practolaemus and Kneria, and a stained and partially cleared skeleton of Kneria. Whole specimens of all genera mentioned have been examined. For the osteology of Cromeria, reliance has had to be placed on the paper by Swinnerton (1903). Ridewood's accounts of the head skeletons of Phractolaemus (190501) and Gonorhynchus ( 1905^) have also been used. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 355 nately Ridewood studied only the head skeletons of a limited number of isospondy- lous groups. Regan (19130) dealt briefly with the classification of the salmonoid fishes and in 1929 presented a synopsis of the full order. Berg's (1940) volume introduced some rather drastic innovations. Chapman (1934-1948) has written a number of valuable papers on the relationships of individual members of the group, and Norman (1957) gave a key to the suborders, families, and genera. Besides the above and many other papers dealing more or less directly with isospondylous classification, a large number of works bear indirectly on the subject. For example, Ford (1937) took up variations in the vertebral column. Whitehouse (19100 and b), Regan (19100 and b), Hollister (1936, 1940), and Schaeffer (1949 : 13, 14) have taken up the caudal skeleton. Starks (1930) wrote the most complete account of the pectoral girdle. Frost (1925) took up isospondylous otoliths. Allis (1904), Wohlfahrt (1937), and Nybelin (1957) have dealt with the lateral line system of the head. Van Dobben (1935), Hofer (1945), Tchernavin (1953), Giinther and Diagram 3 Deckert (1953, etc.) and Kirkhoff (1958) have taken up various aspects of the func- tional morphology of the head. As to soft anatomy, Bridge (1900), de Beaufort (1909), Svetovidov (1950, etc.), Jones and Marshall (1952), and O'Connell (1955) investigated the air bladder. Boas (1880) and Senior (1907, etc.) reported on the heart. Jacobshagen (1912) and Suyehiro (1942) have made comparative studies of the digestive system . Finally, Hyrtl (1856), Lagler and Kraatz (1945), Kapoor (1954), and Takahashi (1957) have dealt with the epibranchial organ. Despite all of the work that has been done, the classification of the isospondylous fishes (Order Clupeiformes) remains unsatisfactory. Indeed, Berg (1940 : 417) has written : " This order, from which a series of higher orders has arisen, represents an artificial assemblage, its separate members, as may be seen from the diagnosis, differing greatly from one another. In time the Clupeiformes will be, doubtlessly, divided in many orders ". Berg's view, I think, stems at least in part from a tendency to look at isospondylous classification at a single time level, for example AB of Diagram 3, whereas if a time dimension, CD, were incorporated into the concept, the " artifical assemblage " aspect would seem to largely disappear. In any event there are four main questions that need to be answered with regard to isospondylous classification, (i) What are the interrelationships of the living members? (2) What are the relationships between living and fossil members? (3) What are the limits of the group? (4) How can it be defined? 356 A CLASSIFICATION OF MODERN ISOSPONDYLOUS FISHES Here, no effort will be made to deal with fossil forms. The classification of fossil Clupeiformes is a large and important subject in itself, and one to which the present author can make little contribution as he has no first hand familiarity with fossil material. The classification of modern isospondylous fishes that follows thus attempts to reconstruct phylogenetic relationships from a knowledge of fishes at only one time level and is consequently prone to all the pitfalls inherent in such a procedure. Yet there are certain points to be made in justification of such a classification. One is that it provides a corpus for the paleontologists to work from, with, or even against. More important, however, is the fact that ichthyologists working with modern fishes are obligated to do as much as they can with the fishes around them. There is virtually no chance that neotenic fishes like Cromeria and Lovettia will ever turn up in the fossil record. Indeed one is inclined to wonder whether the fossil record would not give a very peculiar picture of isospondylous classification because of the overrepresentation of some groups and the underrepresentation of others. Thus the salmonids are usually stream dwellers and are unlikely candidates for fossilization ; the same might be said of the oceanic stomiatoids. Again, some of the most peculiar of the isospondylous fishes are from African fresh waters, and pre-Pliocene fish remains from that continent are as yet few. On the other hand certain of the isospondylous fishes, notably the herring-like forms, seem frequently to die in areas where they fossilize in relative abundance. Thus if the major groups of the paleontologist, e.g., the chirocentrids, do not always correspond to the major groups of the modern ichthyologist, that may merely mean a difference in viewpoint rather than an error in interpretation. Eventually, of course, a single classification of isospondylous fishes based on both fossil and recent material will have to be developed, but the time to attempt this hardly seems to have arrived. As to a hard and fast definition of the Clupeiformes, that, as might be expected of any group so diverse, is impossible except in negative terms. Apparently about the best that can be done is as follows : Caudal fin homocercal. No true spines in the fins. Pelvic fins abdominal and with more than 5 rays (although in the Notopteridae the pelvics are absent or rudimentary and few-rayed, and the Notopteridae, Pantodon and Macristium have subthoracic pelvics). Maxillary usually included in the gape. Scales, if present, cycloid (except Gonorhynchus and certain species of argentinids and osmerids). Air bladder, if present, physostomous (except Lovettia, stomiatoids, and Argentinoidae). Orbitos- phenoid and mesocoracoid generally present ; branched caudal rays most frequently 17. No Weberian ossicles. Lateral line, if present, not running low on sides. Snout not tubular. No electric organs. With regard to the limits of the Clupeiformes, both the " upper " and the " lower " borders of the order have been the subject of controversy. Thus Saint-Seine (1949) removed Flops from the isospondylous fishes and placed it in the Halecostomi, and Nybelin (1957) has stated that Flops is as much a holostean as Amia or Lepiso- steus. Here, Flops will be retained in the Clupeiformes, with the full realization that it becomes extremely difficult to place a lower limit on the order when fossil material is taken into account. A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 357 At the " upper " limit there are six groups of fishes that have been at times included in and at other times excluded from the Clupeiformes by modern authors. That the Anotopteridae, included " incertae sedis " by Berg (1940 : 435), belongs among the iniomous fishes has been adequately shown by Maul (1946) and Nybelin (1946). Berg (1940 : 436) creates a separate order for the Galaxiidae, but this family may apparently be returned to its usual place in the Clupeiformes (see Section 2). The Bathyclupeidae has been juggled about by various authors, sometimes landing in the isospondylous fishes, but it belongs with the percoids where Regan (19136) placed it (N. B. Marshall). The three remaining groups are more controversial. The gymnarchid-mormyrid stock can be traced back with considerable certainty into the osteoglossid-notopterid group of isospondylous fishes. Nevertheless the Mormyriformes seems to show sufficient specializations e.g., in the electric organ and the brain to justify separation at the ordinal level. The controversy regarding the iniomous fishes has a somewhat different basis. Their origin in the Clupeiformes cannot be traced by more than guesswork. However, the one character usually used to differentiate the iniomous from the isospondylous fishes is the exclusion of the maxillary from the gape in the former, and Chapman (1944) and others have pointed out that various isospondylous families also have the maxillary excluded. Thus the differentiation between the two orders breaks down, and Schultz and Stern (1948) have consequently merged them. Nevertheless, it does not seem justifiable to unite two groups just because all themembers of one cannot be separated from all the members of the other by a single character. In the instance of the Mycto- phiformes and the Clupeiformes it is not even a question of where to draw the line in a continuous lineage, for the isospondylous fishes with the maxillary excluded from the gape (except possibly Albula) undoubtedly did not give rise to the iniomous fishes. But if the argument for merging the two orders seems weak, that for keeping them separate as is done here has, in this author's opinion, never been very convinc- ingly presented. (The closest approach seems to be that of Marshall, 1955 : 305). Finally, that the haplomous fishes belong in the Clupeiformes has been convincingly argued by Berg (1936, 1940), and further evidence for this allocation has already been presented in Section 2 of this paper. Order Clupeiformes (Isospondyli) Division I Clupei Suborder I Elopoidei Superfamily I Elopoidae Superfamily II Albuloidae Suborder II Clupeoidei Superfamily I Alepocephaloidae Superfamily II Clupeoidae Suborder III Gonorhynchoidei Superfamily I Chanoidae Superfamily II Gonorhynchoidae Superfamily III Phractolaemoidae Superfamily IV Knerioidae Superfamily V Cromerioidae 358 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES Suborder IV Stomiatoidei Superfamily I Gonostomoidae Superfamily II Stomiatoidae Superfamily III Astronesthoidae Suborder V Salmonoidei Superfamily I Salmonoidae Superfamily II Argentinoidae (Opisthoproctoidae) Superfamily III Osmeroidae Suborder VI Esocoidei (Haplomi) Superfamily I Dallioidae Superfamily II Umbroidae Superfamily III Esocoidae Division II Osteoglossi Suborder VII Osteoglossoidei Superfamily I Hiodontoidae Superfamily II Notopteroidae Superfamily III Osteoglossoidae Certain final points regarding the classification of the isospondylous fishes pre- sented above may be noted. First, N. B. Marshall has called my attention to the omission of three families of dubious systematic position : the Bathylaconidae (Parr, 1948), the Rosauridae (Tucker, 1954) and the Macristiidae (Regan, 1911). These are all families of oceanic fishes based on one or a few specimens. Any attempt to place them in one position or another within the Clupeiformes at the present time would be gratuitous, and it seems better to leave them incertae sedis. As to the better- represented groups, the classification of the superfamilies of Stomiatoidei and Eso- coidei stand as they appear in Berg (1940), and the Superfamily Alepocephaloidae is left as usual in the Clupeoidei. The author has almost no first hand knowledge of any of these groups. Second, as a matter of practice this author believes that orders and suborders should be monophyletic but rather broadly interpreted. In contrast with Berg (1940), he would be very chary about the erection of orders and suborders for a few aberrant forms ; these he would in general relegate to superfamilies (if such a higher category is needed). The reasoning behind this is that superfamilies are usually used by specialists, to whom the names of small divergent groups are significant, but are rarely incorporated in general classifications, where such names would generally have little meaning. Finally, there remains the question of whether the Clupeiformes is polyphyletic. The evidence of the caudal skeleton would appear to be against such a conclusion. The caudal structures of the basal forms Elops, Salmo, Esox, and Hiodon are not only strikingly similar (Section i), they are also widely different from anything among the recent members of the Holostei. Some of the ways in which they resemble each other and differ from the caudal skeletons of lower fishes seem to be of a type that would hardly have been developed independently twice, e.g., the three and only three hypurals to the lower caudal lobe, the peculiarities in the basal articulations A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 359 of these three hypurals (No. i with the terminal vertebra, and Nos. 2 and 3 with the first postterminal centrum), the long uroneurals, etc. Indeed if there is any indication of polyphyletic origin in the caudal structure of the Clupeiformes it would be between the osteoglossid-notopterid stock on the one hand and Hiodon with all the rest of the isospondylous fishes on the other. But such a split separating Hiodon from the osteoglossid groups seems untenable on any other ground than the caudal skeleton. When all features are taken into consideration only two stocks (Divisions) seem distinctly separable within the order. One of these (Division II) contains today a small number of relict forms that are all, in their peculiar fashion, mixtures of primitive and specialized characters. To exemplify the basic nature of the stock no one fish can be selected ; rather it is necessary to give a composite description based on the primitive features found in most of the Hiodontoidae, Osteoglossoidae, and Notopteroidae. If this is done it may be said that in Division II the parasphenoid remains primitive from three points of view. First, it never reaches the rear of the basioccipital ; second, it sometimes retains a pair of projections for articulation with the suspensorium ; and finally, it generally bears teeth. The last two features play a role in the general dentition and palatal construction in the stock as a whole ; in these Division II appears to have remained primitive, quite in contrast to the evolution that has occurred in Division I. Throughout Division II, except Heterotis, the primary portion of the bite is between the parasphenoid and the tongue (cf., Hofer, 1945). Such bony articulations as occur on the top of the mouth are between the parasphenoid and the mesopterygoid, not from the ethmoid-prefrontal area via the palatines to the maxillary pedicels. Indeed, in Division II the palatines end anteriorly in a simple point and the maxillaries merely appear to be toothed cheek bones that have come loose posteriorly ; they have no pedicels. However, if supramaxillaries are a primitive character, Division II has lost them. Also, in the vertebral column Division II would seem to have advanced farther along the road toward the fusion of elements than Division I, and in the caudal skeleton (aside from Hiodon) Division II seems to have evolved in a totally different way from Division I. By contrast Division I is represented today by some primitive forms, notably the elopoids, that in most respects might stand as the ancestors of the whole division. Actually, Flops and Albula in certain characteristics, especially dentition, seem to form better prototypes for the stock than the Jurassic Leptolepis bronni (Rayner, 1937), for in dentition L. bronni seems already to have developed much of the speci- alization of the modern Clupeidae. Returning to modern forms, the Elopoidae and Albuloidae are the only supe'rfamilies in the Division that have retained a para- sphenoid dentition. In the others it has been totally lost, and when teeth reappear along the center of the mouth roof as in Esox they are on a backward projection of the vomer. The parasphenoid-mesopterygoid articulation is also lost in all modern forms though it is present in Leptolepis bronni (Rayner, 1937). On the other hand, a complicated cranium-palatine-maxillary articulation is usually present, giving the whole upper jaw structure an aspect rather similar to that of the higher fishes. An important question regarding Division I is whether it contains within itself two or more independent lineages that are of equal value with Division II. One 360 A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES such possible independent group is the Gonorhynchidae. If, however, Gonorhynchus is related to Chanos, as is suggested in Section 3 of this paper, then its pertinence to Division I would seem fairly secure. A more moot group, in this author's opinion, is the whole salmonoid-haplomous stock. There are only two items known to the author that might serve as an indication of a salmonoid-clupeoid relationship. One is that Albula, which is usually assumed to be related to the clupeoids, seems to show some similarity in the caudal skeleton to that typical of the salmonoids. The other is that the stomiatoids appear to be intermediate in certain respects between the salmonoids and the clupeoids : on the one hand some of the stomiatoids have the two supramaxillaries of the clupeoid fishes and Flops ; on the other, many stomiatoids have the typical salmonoid adipose fin. Especially because of this puzzling link provided by the stomiatoids, it seems best to include the salmonoids in the same division with the clupeoids for the moment. SYNOPSIS OF MODERN CLUPEIFORM FISHES TO SUPERFAMILY i a. Parasphenoid extending nearly to the posterior end of the basioccipital or beyond. Parasphenoid never with lateral processes for articulation with the mesoptery- goid ; anterior ribs usually articulating with small parapophyses that are wedged into pits in the centra ; branched caudal rays often 17. Division Clupei. 2a. Teeth present on the parasphenoid ; a leptocephalous larval form. Suborder ELOPOIDEI 3a. A gular plate . . ."' . . . . . Superfamily ELOPOIDAE 3b. No gular plate ........ Superfamily ALBULOIDAE 2b. No teeth on the parasphenoid ; no leptocephalous larval form. 4a. Lower intermuscular bones usually present ; photophores, if present, not in two rows along lower sides ; posterior neural and hemal spines without antero-posteriorly expanded blades ; no proethmoids. 5a. Supramaxillaries usually present ; gill covers free from the isthmus ; branchio- stegal rays 7 or more ....... Suborder CLUPEOIDEI 6a. Two postterminal centra ; no temporal foramen ; black, deep-water fishes Superfamily ALEPOCEPHALOIDAE 6b. One or no postterminal centra ; a temporal foramen ; silvery, surface-living fishes ......... Superfamily CLUPEOIDAE 5b. No supramaxillaries ; gill covers attached to the isthmus ; branchiostegal rays 3 or 4 . . . . . . . Suborder GONORHYNCHOIDEI 7a. Scales cycloid ; gill openings not restricted ; anterior ribs articulating in part with parapophyses wedged into pits in the centra ; mouth terminal Superfamily CHANOIDAE yb. Scales ctenoid ; gill openings not restricted ; anterior ribs articulating with strong lateral processes from the centra ; mouth inferior Superfamily GONORHYNCHOIDAE 7C Scales cycloid ; gill openings little restricted ; anterior ribs articulating directly with the centra ; mouth superior . . . Superfamily PHRACTOLAEMOIDAE 7d. Scales cycloid ; gill openings restricted ; anterior ribs articulating directly with the centra ; mouth inferior ..... Superfamily KNERIOIDAE 76. No scales ; gill openings restricted ; mouth inferior . Superfamily CROMERIOIDAE 4b. No lower intermuscular bones ; two rows of photophores along lower sides ; posterior neural and hemal spines without antero-posteriorly expanded blades ; no proethmoids ....... Suborder STOMIATOIDEI A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 361 8a. Premaxillary without anterior expansion ; maxillary with two supramaxillaries ; gill arches with gill rakers ..... Superfamily GONOSTOMOIDAE 8b. Premaxillary with a strong anterior expansion extending upward over the ethmoid; maxillary without supramaxillaries ; gill arches without normal gill rakers. ga. Supracleithrum absent ....... Superfamily STOMIATOIDAE 9b. Supracleithrum present ..... Superfamily ASTRONESTHOIDAE 4C. No lower intermuscular bones ; no photophores ; posterior neural and hemal spines with antero-posteriorly expanded blades (except in Lovettia and the Salangidae) ; proethmoids sometimes present .... Suborder SALMONOIDEI loa. Two upturned postterminal vertebral centra . . . Superfamily SALMONOIDAE lob. Not more than one upturned postterminal vertebral centrum. 1 1 a. Oceanic fishes with small pelagic eggs. . . . Superfamily ARGENTINOIDAE. nb. Inshore or freshwater fishes with large, demersal eggs . Superfamily OSMEROIDAE 4d. No lower intermuscular bones, photophores, or antero-posteriorly expanded blades on the posterior neural and hemal spines; paired proethmoids present Suborder ESOCOIDEI I2a. Snout not greatly produced ; no canine teeth. i3a. Scapula, coracoid, and pectoral radials not ossified . . Superfamily DALLIOIDAE I3b. Scapula, coracoid, and pectoral radials ossified . . Superfamily UMBROIDAE i2b. Snout greatly produced ; canine teeth present. . . Superfamily ESOCOIDAE ib. Parasphenoid terminating well short of the posterior end of the basioccipital. Parasphenoid sometimes with lateral processes for articulation with the meso- pterygoid ; anterior ribs either articulating directly with centra or with strong parapophyses that are fused to the centra ; branched caudal rays 16 or fewer Division OSTEOGLOSSI I4a. Three or four uroneurals in the caudal skeleton ; no parasphenoid facet for articulation with the mesopterygoid ; nasals small ; symplectic a normal, splint- like bone ; basisphenoid absent ; top of skull without longitudinal ridges. Superfamily HIODONTOIDAE. I4b. No separate uroneurals nor parasphenoid facet ; nasals enlarged ; symplectic expanded ; basisphenoid present ; top of skull with longitudinal ridges Superfamily NOTOPTEROIDAE I4C. No separate uroneurals ; parasphenoid facets present ; nasals enlarged ; basi- sphenoid present ; top of skull rugose but without longitudinal ridges Superfamily OSTEOGLOSSOIDAE REFERENCES Numbers in parentheses after each item refer to the section of the present paper in which this item was cited. ALLIS, E. P. 1904. The latero-sensory canals and related bones in fishes. Intern. Monats. Anal. Physiol., 4 : 401-502, pis. 8-20. (4) BARRINGTON, E. J. W. 1937. The structure and development of the tail in the plaice (Pleuro- nectes platessa) and the cod (Gadus morrhua). Quart. Jour. Micr. Sci., London, new series, 79 : 447-469, 25 figs, (i) DE BEAUFORT, L. F. 1909. Die Schwimmblase der Malacopterygier. Morph. Jahrb., 39 : 529- 644, i pi., 10 figs. (4) BERG, L. S. 1936. The suborder Esocoidei (Pisces). Bull. Inst. rech. biol. Perm, 10 : 385-391. (In Russian and English). (2, 4) 1940. Classification of fishes, both recent and fossil. Trav. Inst. zool. Acad. Sci. U.R.S.S., Leningrad, 5 : 87-517, 190 figs. (In Russian and English). (2, 3, 4) 1955- The same. 2nd Ed. Ibid., 20 : 1-286, 263 figs. (In Russian). (2) BERTELSEN, E. 1958. The argentinoid fish Xenophthalmichthys danae Regan 1925. Dana- Report No. 45, pp. 3-8, 6 figs. (2) 362 A CLASSIFICATION OF MODERN ISOSPONDYLOUS FISHES BLACKBURN, M. 1950. The Tasmanian whitebait, Lovettia seali (Johnston), and the whitebait fishery. Australian Jour. Mar. and Freshw. Res., 1 : 155-198, 6 pis., 5 figs. (2) BLANCO, G. J. 1935. The development of the homocercal caudal of the blue perch, Taeniotoca lateralis Agassiz. Philippine Jour. Sci., Manila, 56 : 41-44, i pi. (i) BOAS, J. E. V. 1880. t)ber den Conus arteriosus bei Butirinus und anderen Knochenfischen. Morph. Jahrb., 6 : 527-531, i pi. (4) BRIDGE, T. W. 1900. The air-bladder and its connection with the auditory organ in Notop- terus borneensis. Jour. Linnaean Soc. London, Zool., 27 : 503-540, 2 pis. (4) CHAPMAN, W. M. 1934. The osteology of the haplomous fish, Novumbra hubbsi Schultz, with comparative notes on related species. Jour. Morph., 56 : 371-405, 8 figs. (2, 4) 1941. The osteology and relationships of the isospondylous fish Plecoglossus altivelis Temminck and Schlegel. Ibid., 68 : 425-455, 12 figs. (4) 19420. The osteology and relationships of the Argentinidae, a family of oceanic fishes. Jour. Washington Acad. Sci., 34 : 104-117, 8 figs. (2, 4) 19426. The osteology and relationships of the bathypelagic fish Macropinna microstoma, with notes on its visceral anatomy. Ann. Mag. Nat. Hist., ser. n, 9 : 272-304, 9 figs, (i, 2,4) 1943- The osteology and relationships of the bathypelagic fishes of the genus Bathylagus Giinther with notes on the systematic position of Leuroglossus stilbius Gilbert and Thero- bromus callorhinus Lucas. Jour. Washington Acad. Sci., 33 : 147-160, 8 figs. (2, 4) 1944- On the osteology and relationships of the South American fish, Aplochiton zebra Jenyns. Jour. Morph., 75 : 149-165, 10 figs, (i, 2, 4) 1948. The osteology and relationships of the Microstomidae, a family of oceanic fishes. Proc. California Acad. Sci., ser. 4, 26 : 1-22, 12 figs, (i, 2, 4) COHEN, D. M. 1958. A revision of the fishes of the subfamily Argentininae. Bull. Florida State Mus., 3 : 93-173, 8 figs. (2) DEVILLERS, C. 1958. Le crane des poissons. In Grasse, Traite de Zoologie, vol. 13, fasc i, pp. 551-587, figs. 345-443- Paris, (i) EIGENMANN, C. H. 1927. The freshwater fishes of Chile. Mem. Nat. Acad. Sci., 22 : no. 2, 63 pp., 16 pis. (2) EMELIANOV, S. 1935. Die Morphologic der Fischrippen. Zool. Jahrb., Abt. Anat., 60 : 133-262, 53 figs- (2) FORD, E. 1937. Vertebral variation in teleostean fishes. Jour. Mar. Biol. Ass., new ser., 22 : 1-60, pis. 1-16, 18 figs. (4) FROST, G. A. 1925. A comparative study of the otoliths of neopterygian fishes. Ann. Mag. Nat. Hist., ser. 9, 15 : 152-163, pis. 11-13. (4) GARSTANG, W. 1931. The phyletic classification of Teleostei. Proc. Leeds Philos. Lt. Soc., Sci. Sect., 2 : pt. 5, pp. 240-260. (2) GOODRICH, E. S. 1909. Cyclostomes and Fishes. Vol. 9 of Lankester's " Treatise on Zoology ". Adam and Charles Black, London : xvi + 518 pp., 515 figs, (i) GREENE, C. W. & GREENE, C. H. 1914. The skeletal musculature of the king salmon. Bull. U.S. Bur. Fish, for 1913, 33 : 21-60, pis. i, 2, 14 figs, (i) GTJNTHER, K. & DECKERT, K. 1953. Morphologisch-anatomische und vergleichend okolo- gische Untersuchungen iiber die Leistungen des Viszeralapparates bei Tiefseefischen der Gattung Cyclothone (Teleostei, Isospondyli). Zeitschr. Morph. Okol., 42 : 1-66, 25 figs. (4) HINDERSSON, H. A. 1910. tiber die Schwanzflossenmuskulatur der Teleostier. Anat. Anz., 36 : 465-471, 5 figs, (i) HOFER, H. 1945. Zur Kenntnis der Suspensionsformen des Kieferbogens und deren Zusam- menhange mit dem Bau des knochernen Gaumens und mit der Kinetik des Schadels bei den Knochenfischen. Zool. Jahrb., Abt. Anat., 69 : 321-404, 21 figs. (4) HOLLISTER, G. 1936. Caudal skeleton of Bermuda shallow water fishes. I. Order Isospondyli : Elopidae, Megalopidae, Albulidae, Clupeidae, Dussumieriidae, Engraulidae. Zoologica, New York, 21 : 257-290, 53 figs, (i, 4) A CLASSIFICATION OF MODERN ISOSPOND YLOUS FISHES 363 HOLLISTER, R. G. 1940. Young Megalops cyprinoides from Batavia, Dutch East India, includ- ing a study of the caudal skeleton and a comparison with the Atlantic species, Tarpon atlanticus. Ibid., 24 : 449-467, 21 figs, (i, 4) HUBBS, C. L. 1953. Synonymy of the bathypelagic fish genus Rhynchohyalus, referred to the expanded family Argentinidae. Copeia, 1953, pp. 96-97. (2) HYRTL, C. J. 1856. Uber die accessorischen Kiemenorgane und den Darmcanal der Clupea- ceen. Denkschr. Akad. Wiss. Wien, math.-nat. Cl., 10 : 47-58, 3 pis. (4) JACOBSHAGEN, E. 1912. Untersuchungen iiber das Darmsystem der Fische und Dipnoer. Teil II. Materialien zur vergleichenden Anatomic des Darmkanals der Teleostomen nebst einer einleitenden Ubersicht. Jenaische Zeitschr. Naturwiss., 49 : 373-810, pi. 18, 164 figs. (4) JONES, F. R., & MARSHALL, N. B. 1953. The structure and functions of the teleostean swim- bladder. Biol. Rev., 28 : 16-83, 7 figs. (4) KAPOOR, B. G. 1954. The pharyngeal organ and its associated structures in the milk-fish, Chanos chanos (Forskal). Jour. Zool. Soc. India, 6 : 51-58, 5 figs. (4) KENDALL, W. C. 1921. Peritoneal membranes, ovaries, and oviducts of salmonoid fishes and their significance in fish-cultural practices. Bull. Bur. Fish., Washington, 37 : 185-208, ii figs. (2) 1935. The fishes of New England. The Salmon family. Part 2 The Salmons. Mem. Boston. Soc. Nat. Hist., 9 : 1-166, n col. pis. (2) KENDALL, W. C. & CRAWFORD, D. R. 1922. Notice of a spiral valve in the teleostean fish Argentina silus, with a discussion of some skeletal and other characters. Jour. Washington Acad. Sci., 12 : 8-19, 2 figs. (2) KIRKHOFF, H. 1958. Functionell-anatomische Untersuchung des Visceralapparates von Clupea harengus L. Zool. Jahrb., Abt. Anat., 76 : 461-570, 63 figs. KOLLIKER, A. 1860. Uber das Ende der Wirbelsaule der Ganoiden und einiger Teleostier. Leipzig : 27 pp., 4 pis. (i) LAGLER, K. F. & KRAATZ, W. C. 1945. Pharyngeal pockets in the gizzard shad, Dorosoma cepedianum (Lesueur). Pap. Michigan Acad. Sci., Arts, Lett., 30 : 311-320, 2 pis., i fig. (4). Loxz, T. 1864. Uber den Bau der Schwanzevirbelsaule der Salmoniden, Cyprinoiden, Per- coiden und Cataphracten. Zeitschr. Wiss. Zool., 14 : 81-106, 4 pis., 2 figs, (i) MARSHALL, N. B. 1955. Alepisauroid fishes. Discovery Repts., 27 : 303-336, pi. 19, 9 figs. (4) MAUL, G. E. 1946. Monografia dos peixes do Museu Municipal do Funchal. Ordem Iniomi. Bol. Mus. Mun. Funchal, No. 2, art. 2, pp. 5-61, 22 figs. (4) NORMAN, J. R. 1957. A draft synopsis of the orders, families and genera of recent fishes and fish -like vertebrates. British Museum (Natural History), London : 649 pp. (4) NYBELIN, O. 1946. Sur quelques especes du genre Anotopterus Zugmayer et sur la position systematique de la famille des Anotopteridae. Goteborgs Vetensk. Samh. Handl., ser. 6, SB : 3-16, i pi. (4) 1957- L 63 canaux sensoriels du museau chez Flops saurus (L.) Notice preliminaire. Arkiv for Zoologi, 10 : 453-458, 3 figs. (4) O'CONNELL, C. P. 1955. The gas bladder and its relation to the inner ear in Sardinops caerulea and Engraulis mordax. Bull. U.S. Fish Wildlife Serv., 56 : 503-553, 15 figs. (4) PARR, A. E. 1937. Concluding report on fishes. Bull. Bingham Oceanogr. Coll., 3, art. 7, 79 pp., 22 figs. (2) - 1948. The classification of the fishes of the genera Bathylaco and Macromastax, possible intermediates between the Isospondyli and the Iniomi. Copeia, 1948, pp. 48-54, 2 figs. (4) RAMANUJAM, S. G. M. 1929, The study of the development of the vertebral column in teleosts, as shown by the life history of the herring. Proc. Zool. Soc. London, 1929, pp. 365-414, 28 figs, (i) RAYNER, D. N. 1937. On Leptolepis bronni Agassiz. Ann. Mag. Nat. Hist., ser. 10, 19 : 46-74, 14 figs. (4). REGAN, C. T. 1905. A revision of the fishes of the family Galaxiidae. Proc. Zool. Soc. London 1905, pp. 363-384. 4 Pis- (2) 364 A CLASSIFICATION OF MODERN ISOSPOND YLOU S FISHES REGAN, C. T. igioa. The caudal fin of the Elopidae and of some other teleostean fishes. Ann. Mag. Nat. Hist., ser. 8, 5 : 354-358, 2 figs, (i, 4) - 19106. On the caudal fin of the Clupeidae, and on the teleostean urostyle. Ibid., pp. 531-533. 2 figs, (i, 4) 1911. On the systematic position of Macristium chavesi. Ibid., 7 : 204-205, i fig. (4) 19130. The antarctic fishes of the Scottish National Antarctic Expedition. Trans. Roy. Soc. Edinburgh, 49 : 229-291, n pis., 6 figs. (2, 4) 19136- The classification of the percoid fishes. Ann. Mag. Nat. Hist., ser. 12,8: 111-145. (4) 1929. " Fishes " in Encyclopaedia Britannica, ed. 14, 9 : 305-328. (2) RIDEWOOD, W. G. 19040. On the cranial osteology of the fishes of the families Elopidae and Albulidae, with remarks on the morphology of the skull in the lower teleostean fishes generally. Proc. Zool. Soc. London, 1904, pt. 2, pp. 35-81, figs. 8-18. (4) 19046. On the cranial osteology of the fishes of the families Mormyridae, Notopteridae and Hyodontidae. Jour. Linnaean Soc. London, Zool., 29 : 188-217, pis. 22-25. (4) - 19050. On the cranial osteology of the clupeoid fishes. Proc. Zool. Soc. London, 1904, pt. 2, pp. 448-493, figs. 118-143. (4) 19056. On the cranial osteology of the fishes of the families Osteoglossidae, Pantodontidae and Phractolaemidae. Jour. Linnaean Soc. London, Zool., 29 : 252-282, pis. 30-32. (4) 19056. On the skull of Gonorhynchus greyi. Ann. Mag. Nat. Hist., ser. 7, 15 : 361-372, pi. 16, i fig. (3, 4) SAINT-SEINE, P. DE. 1949. Les poissons des calcaires lithographiques de Cerin (Ain). Nouv. Arch. Mus. Hist. nat. Lyon, Fasc. 2, pp. 1-357, 26 P^ s -. I2 fig 3 - (4) SCHAEFFER, B. 1949. A teleost from the Livingston formation of Montana. American Mus. Novitates No. 1427, 15 pp., 2 figs. (4) SCHMALHAUSEN, I. I. 1913. Zur Morphologic der unpaaren Elossen. II. Bau und Phylogenese der unpaaren Flossen und insbesonders der Schwanzflosse der Fische. Zeitschr. Wiss. Zool., 104 : 1-80, 14 figs. SCHULTZ, L. P. & STERN, E. M. 1948. The ways of fishes. 1). Van Nostrand, New York : 264 pp. (2, 4) SCOTT, E. O. G. 1936. Observations on fishes of the family Galaxiidae. Pt. i. Pap. Proc. Roy. Soc. Tasmania, 1935, pp. 85-112, 4 figs. (2) SENIOR, H. D. 1907. Note on the conus arteriosus of Megalops cyprinoides (Broussonet). Biol. Bull., 12 : 378-379, 2 figs. (4) STARKS, E. C. 1930. The primary shoulder girdle of the bony fishes. Stanford Univ. Ser., Biol. Sci., 6 : 149-239, 38 figs. (4) STOKELL, G. 1941. A revision of the genus Retropinna. Rec. Canterbury Mus., 7 : 361-372, Pis. 55-57- (2) I945- The systematic arrangement of the New Zealand Galaxiidae. Part I. Generic and sub-generic classification. Trans. Roy. Soc. New Zealand, 77 : 124-137, pis. 8-12. (2) I955- Fresh water fishes of New Zealand. Simpson and Williams, Christchurch : 145 pp., 42 pis. (2) SUYEHIRO, Y. 1942. A study of the digestive system and feeding habits of fish. Japan. Jour. Zool., 10 : 1-303, 15 pis., 190 figs. (4) SVETOVIDOV, A. N. 1950. Peculiarities in the structure of the air-bladder in Clupeidae. C. R. Acad. Sci. Moscou, n. s., 74 : 643-644. (In Russian). (4) SWINNERTON, H. H. 1903. The osteology of Cromeria nilotica and Galaxias attenuatus. Zool. Jahrb., Abt. Anat., 18 : 58-70, 16 figs. (2) TAKAHASHI, N. 1957. O n the so-called accessory respiratory organ " gill helix " found in some clupeiform fishes, with special reference to the function and genealogy. Japan. Jour. Ichthy., 5 : 71-77, i pi. (4) TCHERNAVIN, V. V. 1953- The feeding mechanisms of a deep sea fish Chauliodus sloani Schneider. British Museum (Natural History), London : 99 pp., 10 pis., 36 figs. (4) TOTTON, A. K. 1914. The structure and development of the caudal skeleton of the teleostean fish, Pleuragramma antarcticum. Proc. Zool. Soc. London, 1914, pp. 251-262, 2 pis. (i) A CLASSIFICATION OF MODERN I SOSPOND YLOUS FISHES 365 TREWAVAS, E. 1933. On the structure of two oceanic fishes, Cyema atrum Gimther and Opisthoproctus soleatus Vaillant. Proc. Zool. Soc. London, 1933, pp. 601-614, 2 pl s -> 8 n 8 s - (1,2) TUCKER, D. W. 1954. Tne " Rosaura " Expedition. 4. Report on the fishes . . . Part I. Bull. British Mus. (Nat. Hist.), Zool., 2 : 163-214, pis. 7, 8, 19 figs. (4) VAN DOBBEN, W. H. 1935- Uber den Kiefermechanismus der Knochenfische. Arch. Neerland. Zool., 2 : 1-72, 50 figs. (4) VOIGT, E. 1934. Die Fische aus der mitteleozanen Braunkohle des Geiseltales. Nova Acta Leopoldina, n. F., 2 : 22-146, 14 pis., 23 figs. (2) WHITEHOUSE, R. H. 19100. The caudal fin of the Teleostomi. Proc. Zool. Soc. London, 1910, pp. 590-625, pis. 47-50, fig. 57. (i, 4) 19106. Some remarks on the teleostean caudal fin. Ann. Mag. Nat. Hist., ser. 8, 5 : 425- 427- (i, 4) 1918. The evolution of the caudal fins of fishes. Rec. Indian Mus., 15 : 135-142. (i) WILIMOVSKY, N. J. 1951. The classification of isospondylous fishes. A. S. I. H. Committee on Fish Classification, Circular No. 4, 38 pp. (Mimeographed). (2, 3, 4) WOHLFAHRT, T. A. 1937- Anatomische Untersuchungen iiber die Seitenkanale der Sardine (Clupea pilchardus Walb.). Zeitschr. Morph. Okol. Tiere, 33 : 381-411, 15 figs. (4) THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS (Based on a collection made in 1957) B. P. HALL BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 7 LONDON: 1960 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS (Based on a collection made in 1957) BY B. P. HALL 367-453 ; Plates 5-6 ; 2 Text-figures BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY) ZOOLOGY Vol. 6 No. 7 LONDON: 1960 THE BULLETIN OF THE BRITISH MUSEUM (NATURAL HISTORY), instituted in 1949, is issued in five series, corresponding to the Departments of the Museum, and an Historical series. Parts will appear at irregular intervals as they become ready. Volumes will contain about three or four hundred pages, and will not necessarily be completed within one calendar year. This paper is Vol. 6, No. 7 of the Zoological series. Trustees of the British Museum, 1960 PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM Issued July, 1960 Price Forty Shillings THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS (Based on a collection made in 1957) By B. P. HALL CONTENTS Page INTRODUCTION ........... 370 Scope of paper Acknowledgments Deposition of specimens collected ITINERARY ............ 372 VEGETATION ........... 373 Main vegetational regions, with map Vegetation at each Camp TABLES OF SPECIMENS COLLECTED ....... 379 Table I. Birds of the woodlands II. Birds of the forests III. Birds of grassland, reeds and open country .... IV. Birds of the thornbush semi-desert, tangled thickets and rocks of western districts DISTRIBUTION THROUGHOUT AFRICA OF SPECIES FOUND IN ANGOLA . . 390 Woodland birds Forest birds Grassland birds Birds of the western districts AVIFAUNAL ZONES AND SPECIATION ....... 395 Avifaunal zones, with map Speciation Examples of stages in speciation BREEDING SEASONS .......... 401 General discussion Some relevant rainfall figures Non-passerines Woodland passerines Forest passerines Grassland passerines Passerines of western districts NEW RECORDS FOR ANGOLA ......... 405 Species not hitherto recorded in the country Extension of ranges New races which have been described on birds from the collection Sight records of interes SYSTEMATIC AND FIELD NOTES ON SELECTED SPECIES .... 406 Scope of the notes Notes on 84 species BIBLIOGRAPHY . . . . . . . . . . . 450 APPENDIX . . . . . . . . . . . 451 List of Angola place-names mentioned in the text, other than camps ZOOL. 6, 7. 22 370 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS INTRODUCTION SINCE Bocage published the Ornithologie d' Angola in 1881 no major work has appeared on the birds of the country as a whole and even the vast collection amassed by Ansorge in the early part of the century has never been fully written up. Thus little knowledge has been available to the student on the ranges and habitats of the many forms endemic to Angola although the discoveries of such collectors as Rudyerd Boulton, Gerd Heinrich and Rudolf Braun in the past twenty-five years have shown it to be a country of exceptional interest. This interest lies not only in its endemic species but also in that it is a meeting place for the avifauna of West Africa, found in the lowland forests of the west ; of South West Africa, found in the coastal thorn country and desert ; of East Africa, in the woodlands of the centre and east. There are also some representatives of a specialised montane avifauna in the Bailundo highlands. Chapin's Birds of the Belgian Congo has done much to remedy the lack of knowledge but, of necessity, deals only briefly with Angola birds, and only with those that are found also in the Congo. In the British Museum workers have been handicapped by the paucity of specimens from Angola as well as by the lack of literature on the country, and it was with a view to filling some of the gaps in the National Collection and with learning something of the habitats and ecology of the birds, that I decided to make a collecting trip to Angola in 1957. This paper is intended to correlate the birds that were collected with the type of country and vegetation in which they were found. For this purpose all specimens have been listed in four " Tables of Specimens Collected," divided according to their habitats, rather than in a systematic list. The main vegetation regions of the country and the particular vegetation at each camp are described and discussed in relation to the distribution of the species throughout Africa and with regard to their possible bearing on the progress of speciation. Additional notes have been made on breeding seasons and on field habits and systematics where these have proved of particular interest. The usefulness of these discussions is necessarily limited by the lack of personal knowledge of the country outside the route of the expedition and the paper as a whole is intended only to supplement the work that is being done concurrently, chiefly in the United States, on other collections from Angola, so that finally there may be a real understanding of the place of the Angola avifauna in African ornithology. A cknowledgments For the expedition I was fortunate to have the support of the Zoology Depart- ment, British Museum (Natural History), and in being able to get together an Anglo- American party following the Pan-African Ornithological Congress in Livingstone in July 1957. The American members were Mr. and Mrs. Donald S. McChesney, Research Associates of Cornell University. Their chief interest was in photography and sound-recording of bird songs, although they took an active interest in the specimens collected during the time they were with the rest of the expedition : THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 371 this time was unfortunately very limited as private affairs compelled them to return to the States on 2nd September. Mr. John Williams, from the Coryndon Museum, Nairobi, came to take charge of the skinning ; he also had to leave about the same time having been unluckily dogged by ill-health throughout. The chief collectors were Lt.-Gen. Sir Gerald Lathbury, K.C.B., D.S.O., M.B.E., and Mr. A. L. (Tony) Archer from Nairobi, who brought his Land Rover and Owaka and Ali as mechanic and cook. John Williams and I also collected when time allowed from skinning. During their time in Angola Mr. McChesney engaged Mr. Mario Pirelli from Luanda as an interpreter and guide to the country. The trustees of the Percy Sladen and of the Godman Trusts made grants towards my travelling expenses and the British share of the expedition, and a grant was made from the Purchase Grant of the Museum. Further expenses were born by Mr. McChesney, General Lathbury and myself, Mr. McChesney also being responsible for the financing of the American share. The Committee of the Bird Exploration Fund contributed with the loan of a gun and the gift of some ammunition. In Livingstone Mr. Rudyerd Boulton gave valuable advice on the country and habitats of the birds which was of great assistance in planning our route. In Angola we received unfailing help from all Government officials with whom we came in contact, and also helpful advice on the birds of the area from Mr. R. Braun at Sa da Bandeira. Since returning to the Museum I have received great assistance in working out the collection from colleagues in the Bird Room and Mr. A. W. Exell, in the Botany Department, and through loans and letters from many ornithologists including Dr. Dean Amadon, Professor J. Berlioz, Mr. C. W. Benson, Mr. P. A. Clancey, Dr. H. Friedmann, Mr. J. C. Greenway, Mr. R. E. Moreau, Prof. N. E. Collias, Miss M. Patterson, Captain C. R. S. Pitman, Dr. A. L. Rand, Prof. H. Schouteden, Dr. W. Serle, Mr. R. H. N. Smithers, Prof. E. Stresemann, Mr. M. P. Stuart Irwin, Mr. M. A. Traylor, and Mr. C. M. N. White. Mr. Smithers also kindly had some birds skinned for me in Bulawayo which had been preserved in salt. Mr. B. C. Lack, of the Metereological Office, supplied rainfall figures. Mr. C. W. Mackworth-Praed and the late Captain C. H. B. Grant have allowed me full use of the manuscript of the second part of the African Handbook of Birds covering Southern Africa. Dr. G. Rudebeck has supplied me with a list of birds collected on the Visser- Transvaal Museum Expedition to Angola in 1956, and allowed me to examine specimens. Commander A. M. Hughes and Mrs. M. Gull have drawn the maps. To all these people I am very much indebted, but above all to my companions for their contributions in their respective spheres, especially to the two collectors who were responsible for the unusually large variety of birds collected in two months. Deposition of Specimens Collected In all 980 specimens were obtained of 306 species, including the second and third known specimens of Prionops gabela. The majority were skinned on the spot, but some less important were preserved in salt for eventual treatment in the Museum (see Ibis, 1959 : 254), and fifty-four were preserved in spirit as anatomical specimens. 372 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS At the request of the Portuguese authorities a number of duplicate specimens have been sent to the Museum in Luanda. A proportion of the remainder, mainly from eastern districts, are in the American Museum of Natural History, a small number in the National Museum of Southern Rhodesia, Bulawayo, and the bulk in the British Museum. The photographs and sound-tracks have been deposited at Cornell University. ITINERARY The expedition entered Angola from the Belgian Congo along the line of the Ben- guela railway at Dilolo. It had been hoped to do some exploratory work in the little known areas of the east and south-east, but Williams' illness and shortage of time made this impossible. Thus little new ornithological ground was broken except in some parts along the main road from the frontier to Nova Lisboa, and energies were mostly concentrated on seeing as many different types of country as possible, and in filling the known gaps in the British Museum collection from such well- known districts as Mt. Moco, Sa da Bandeira, Chingoroi and Gabela. A full list of camps follows with the altitudes and dates so that these may be correlated with the Tables of Specimens collected at the different camps : Date of arrival Locality of camp Lat. Long. Alt. 27th July . Luau R. Angola/Congo . 10.42 S. . 22.12 E. . 3,600' border 7 m. E. of Texeira de Souza ist August . Luacano . 11.16 S. . 21.38 E. . 3, 550' 3rd ,, . Lake Dilolo . 11.308. . 22.00 E. . 3,500' 6th . Luacano 7th ,, . 28 m. W. of Vila Luso . 11.478. . 19.52 E. . 4,400' 8th . 5 m. W. of Munhango . 12.128. . 18.42 E. . 4,700' gth ,, .5m. W. of Vila General . 12.03 S. . 17-30 E. . 4,800' Machado loth . Vouga . 12. ii S. . 16.47 E. . 5,700' I2th .5m. E. of Nova Lisboa . 12.44 S. . 15-47 E. . 5,500' I4th . Mt. Moco . 12.27 S- . 15.10 E. . 6,000- 8,000' 25th .5111. N. of Vila Flor . 12.41 S. . 15.31 E. . 5,500' 27th ,, . Caconda . 13.43 S. . 15.06 E. . 5,200' 28th ,, . Quipungo . 14.49 S. . 14.34 E. . 4,400' 2Qth . Sa da Bandeira . 14.55 S. . 13.30 E. . 5,600' 3oth . Leba . 15.04 S. . 13.16 E. . 5,800' 4th September . Cahinde (Tampa) . 15.30 S. . 13.27 E. . 2,500' 6th .5m. S.W. of Sa da Ban- . . . 6,600' deira 7th . Chingoroi . 13.37 S. . 14.01 E. . 3,300' nth ,, . 12 m. S.E. of Benguela . 12.358. . 13.25 E. . S.L-5oo' 1 3th ,, . 60 m. N. of Lobito on Novo . . c. 800' Redondo road I4th ,, . 20 m. N.E. of Novo Redon- . . . c. 1,000' do on Gabela road THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 373 Date of arrival 1 5th 20th 25th 2yth Locality of camp Cuvo R. 12 m. S.W. of Gabela 40 m. S. (by road) of Mum- bondo Calulo Quimberima plantation 40 m. S.E. of Calulo Other localities by roadside: gth August . Coemba, 40 m. W. of Mun- hango Saiangikilo R. 16 m. W. of Munhango i3th . Quipeio, 25 m. N.W. of and Nova Lisboa 25th 7th September . Quilengues, on Sa da Ban- deira-Chingoroi road Lat. 10.48 S. 10.09 S. 10.00 S. 12.26 S. 14.00 S. Long. 14.20 E. 14.15 E. M-53 E. 15.30 E. 14.00 E. Alt. 2,500' 1,000' 3.3oo' 3,700' c. 6,000' c. 3,900 VEGETATION Main Vegetational Regions All the main vegetation regions of tropical Africa are represented in Angola with the exception of extensive marshlands of the "Sudd" type. These regions are indicated on the following map which has been adapted from the Vegetation Map of Africa published on behalf of FAssociation pour 1'Etude Taxonomique de la Flore d'Afrique Tropicale with the assistance of U.N.E.S.C.O. (Oxford University Press : 1959), with slightly more detail in those districts covered by the expedition. The regions can be summarized as follows : (a) Savannah Woodland In eastern and central Angola this is largely of the Brachystegia and Isoberlinia type and is extensive, being found even up to 7,000 feet on the mountain slopes. For convenience it will be referred to throughout as Brachystegia woodland. In the west, in a narrow strip along and below the escarpment, there is rather richer vegetation, shown on the map as relatively moist types of savannah woodland. In this strip Brachystegia no longer predominates, there is more variety and Baobabs are common. (b) Grassland Patches of grassland often very flat and poorly drained are interspersed through- out the Brachystegia woodland. In the west there is more open country with many wide stretches of rolling grassland in which other vegetation is confinen to thickets and clumps, mostly in the river and stream valleys. 374 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS VEGETATION MAP OF ANGOLA 100 Montane Forest Miles KEY TO VEGETATION MAP Woodlands and Savannahs with abundant Brachystegia fcscarpment or Riverine Forest Mixed Woodlands and Savannahs relatively moist types Mixed Woodlands and Savannahs, relatively dry types with abun- dant Acacia T. T . T Grass Steppe, Luanda type Desert THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 375 (c) Semi-desert and Thorn In the south and south-west the acacia thornveld of the Kalahari and South West Africa extends into Angola and along the coastal strip to the latitude of Lobito. Here the sandy plain gives way to the rolling grassland and the predominance of acacia diminishes northwards to Luanda. (d) Desert The Namib desert of South West Africa extends in a narrow strip up the coastal plain into southern Angola, merging into the semi-desert or scrub a little north of Mo9amedes and in the foothills towards the base of the Chela escarpment. (e) Forest The forests of Angola, excluding Cabinda in which true tropical rain forest is found, are of three types. True montane forest of great antiquity is found in pockets on the highest mountains of the Bailundo highlands such as Mt. Moco and Mt. Soque. It was also fairly extensive on the Mombolo plateau when Boulton collected there in 1931, but we were informed in Nova Lisboa that it has largely been cut down. Remnants of forest on the Chela escarpment may also prove to be relic montane type. The lowland forests are of more recent origin and can be divided into what might be termed escarpment forest, and gallery or riverine forest. The escarpment forest is found between 1,000 feet and 3,000 feet in sometimes extensive patches below or on the face of the escarpment in areas which combine the high summer rainfall of inland districts with high all-the-year-round humidity and frequent mists due to the meeting of the cold waters of the Benguela current with the warm equatorial waters. Gallery forest is found in patches along the rivers of the north and east, and also along some of the west coast rivers below the escarpment. Vegetation at Camps The expedition traversed all the main vegetation regions with the exception of the true desert and it seems useful to discuss in more detail the country at each camp, and on the road between. The camps and districts discussed have been numbered to correspond with the columns in the Tables of Specimens collected (pages 379-39 1 )- (1) Luau River Most of the collecting was done in the narrow strip of riverine forest or the Brachy- stegia woodland around. There was no open grassland but some clearings in the woodland and the edges of cultivation. (2) Luacano Area There was some riverine forest and Brachystegia woodland, but the camp was chiefly notable for the wide flat treeless plain crossed by damp ditches which extended some miles east and south of Luacano village. 376 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS (3) Lake Dilolo Most of the collecting was done in the grassy swamps and reed-beds on the border of the lake near the Government Rest House, and a little in some woodland a few miles to the west. There was no time to explore the extensive possibilities of the lake, the largest in Angola, or to organize the use of native dugout canoes which would be the most suitable transport for the purpose. (4) Roadside Localities between Luacano and Vouga All camps were made in Brachystegia woodland but this was interspersed with grassy dambos similar to the grassland at Luacano in their flatness and lack of bushes, but not so extensive. (Plate 5, B). (5) Vouga The grasslands here differed in being extensive undulating plains with moist marshy patches in the valleys and some patches of scrubby bushes. A small patch of stunted woodland with some bushes and cultivation surrounded the camp. (6) Nova Lisboa area This area was predominantly Brachystegia woodland with some grassy clearings and cultivation. (7) Foothills of Mt. Moco The lower slopes were covered with Brachystegia woodland with thick tangled grasses in the stream beds, and grassland and cultivation below. (8) Mt. Moco, over 6,000 feet Brachystegia was found up to 7,000 feet on some slopes and montane forest was limited to patches lining precipitous gullies at the head of the streams on the upper slopes, descending to about 6,800 feet on some of the eastern and southern slopes. The forest trees were nowhere very tall or very dense but mostly of ever- green species totally different from the woodland types and with tangled under- growth beneath. The bare slopes above the woodland and away from the forest were rocky and grassy with scattered aloes and small scrub, with numerous plants in flower. (9) Vila Flor, and localities on Luimbale-Nova Lisboa road This area was predominantly Brachystegia woodland with some grassy clearings and cultivation. (10) Caconda and Caconda-Quipungo road This country was similar to Vila Flor until a change was noted about 40 miles south of Caconda where the trees appeared much the same size and form but were notably barer of leaves. This change would not have been worthy of note except that it corresponded with some change in the bird life, four species being met within THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 377 a few miles which had not hitherto been encountered, namely, Neocichla gutturalis, Thamnolea arnotti, Turdoides jardinei and Eurocephalus anguitimens. (n) Quipungo The first patch of thorn country was encountered here, with some grassy clearings and a small patch of Brachystegia. (12) Sa da Bandeim and the roads to Quipungo and Chingoroi This was very mixed country with thorn and Brachystegia alternating, and with richer and more varied vegetation northwards towards Chingoroi as the road descends to 3,000 feet. An overnight camp at Sa da Bandeira was made on top of the escarpment above the town and some collecting was done in the stunted Brachystegia on the slopes and the rocks above. (13) Leba and Humpata-Jau road The camp at Leba was in an area of mixed vegetation. It was situated about a mile inland from the edge of the Chela escarpment at a point where a stream cuts its way through a gorge in a series of waterfalls down 3,000 feet to the foothills and plains at Capangombe. Immediately around the camp was an area of tangled bush and thorn full of aromatic shrubs ; in some clearings by the river were remnants of forest and a patch of forest was found by the river near the foot of the escarpment. The rocks and fringing trees of the escarpment were the typical habitat of Xeno- copyschus ansorgei, and in one place along the edge was an extensive patch of tangled evergreen bush, difficult to penetrate, which contained few birds. Other slopes above the river were covered in stunted Brachystegia, and on the plateau running back from the escarpment patches of Brachystegia were mixed with patches of thorn and occasional grassy or cultivated clearings. (Plate 5, A). In such mixed country it was not always easy to define the habitat of specimens collected and this should be borne in mind when consulting the Tables. Some collecting of larks and pipits was done on open grassland on the Humpata-Jau road, and a party of water-birds was seen at a small lagoon in the centre of these grasslands. (14) Cahinde area Included in this area, is the road down the escarpment where the slopes are lined with stunted Brachystegia and Mopane. This gives way at the bottom of the escarp- ment among the foothills to typical semi-desert country of mixed thorn and Terminalia or Combretum, with the Mopane becoming less common westwards towards the desert. An excursion was made westwards for fifty miles from Cahinde without coming out of the foothills into true desert. (15) Chingoroi The patch of forest in which most of the collecting was done might be described as part riverine and part escarpment being associated with a stream, but more 378 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS extensive than the usual riverine strips. Other birds were collected in reeds and grasses bordering the stream. (16) Benguela area (3-15 Miles South-east] The descent from Chingoroi to the coastal strip behind Benguela was gradual and through mixed country which gave way on the coastal plain to typical thorn bush resembling the Kalahari and the Vornamib. It was more open than the country at Cahinde with more grassland and acacia and no Mopane, and only broken by rocky kopjes. (17) Lobito Novo Redondo-Gabela road The road to Novo Redondo wound over the rolling grasslands mentioned in the discussion on vegetational regions. Thorn bushes and Euphorbia were found on the hill-sides and frequent clumps of dry tangled thickets similar to those found on above the escarpment at Leba. Baobabs were common. Between Novo Redondo and the foot of the escarpment there were clumps of richer vegetation bordering streams and lagoons, but no collecting was done here. (18) Cuvo River, 12 Miles SW. of Gabela All collecting from this camp was done within three or four miles of the river and in the forest on the face of the escarpment, which is again precipitous at this latitude, though not to such a degree as at Leba. The forest is largely underplanted with coffee and the original undergrowth destroyed, but there was an uncleared patch in which we collected not far from the camp. The forest ended abruptly at the top of the escarpment before the town of Gabela and also a few miles further south along the road beyond our camp. The few species listed as collected in grassland in this camp were from tangled grass near the river's edge or a patch of disused cultivation in a forest clearing. (19) 40 Miles South of Mumbondo and road to Luanda Very little collecting was done on the road from Gabela to Luanda via Mumbondo and Muxima, but it was noteworthy for finding Prionops gabela well outside the forest area at the foot of the escarpment in a patch of mixed bush and woodland between cultivated clearings. (20) Calulo and Quimberima (40 Miles SE. of Calulo) After the main expedition was over I collected a few birds at Calulo which is situated in open grassland with some patches of stunted Brachystegia and bush, and on the plantation of Mr. and Mrs. O. Kroel at Quimberima, where they very kindly put me up for two nights. The Quimberima birds were collected in a tiny patch of forest edging a stream and bordered by Brachystegia. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 379 Photographs Photographs and further descriptions of many of the localities and types of vegetation discussed will be found in Lynes' two papers on his expeditions to Angola (1933 and 1938), in Heinrich's recent paper (1958), and in Gosswelier & Mendonca (i939)- TABLES OF SPECIMENS COLLECTED Key to Tables Species which appear in more than one table, or which are referred to in the Systematic and Field Notes, have cross references in the left hand columns. Collecting localities are grouped into districts for estimate of breeding dates and numbered as follows : 1. Luau R. 2. Luacano area : Texeira de Sousa Luacano road. 3. Lake Dilolo. 4. Vila Luso Vouga road. 5. Vouga. 6. Nova Lisboa area. 7. Foothills of Mt. Moco (6,000 feet). 8. Mt. Moco over 6,000 feet. 9. Vila Flor : Quipeio : Nova Lisboa Luimbale road. 10. Caconda : Caconda Quipungo road. 11. Quipungo. 12. Sa da Bandeira : Sa da Bandeira Quipungo road : Sa da Bandeira Chingoroi road. 13. Leba : Humpata Jau road. 14. Cahinde : road down escarpment : Cahinde Mo5amedes road (40 miles). 15. Chingoroi. 1 6. Area 3 12 miles S.E. of Benguela. 17. Lobito Novo Redondo Gabela road. 18. Cuvo R., 12 miles S.E. of Gabela. 19. Forty miles south of Mumbondo. 20. Calulo : Quimberima (40 miles S.E. of Calulo). Unsexed specimens are listed by number only (i.e. i (J, 2 ?, i). Specimens in spirit are listed in brackets. Young birds not fully grown are listed as " juv." Young birds fully grown are listed as " imm." " Summer " breeding season indicates any month between November and May. 380 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS TABLE I. Birds Other references Eastern Districts Central Syst. Tables Notes i 2 3 Breeding 4 5 6 4 Francolinus finschi * 6 coqui angolensis IcJ 15 Chrysococcyx klaasi . (I?) 1 6 Cuculus c. cafer I cJ 19 Poicephalus meyeri damarensis . II 24 Coracias spatulatus IcJ Summer 25 Halcyon chelicuti (I) 28 Merops nubicoides i 6" ? Sept. 29 Melittophaguspusillusmeridionalis 30 Dicrocercus h. hirundineus . icJ (i) Summer 31 Tockus p. pallidirostris to" 34 Glaucidium perlatum licua . 37 Semeiophorus vexillarius 40 Caprimulgus pectoralis fervidus . 3? Summer _7 icJ - 45 Co&us striatus congicus * It ? Aug. 47 castanotus II, IV * 48 Lybius torquatus bocagei . 49 Buccanodon anchietae anchietae . * * i&9 ? July onwards - rex . katangae . i cJ Summer i 6", i ? 52 Tricholaema diadematum fronta- tum 54 Pogonolius chrysoconus extoni IV 2? ? Aug. /Sept. 56 Prodotiscus regulus . i Summer 29 3 31 i 9, 34 Summer 37 40 I (J. 2 9, z July - 45 47 i juv. 6* 3 i$ ? Sept. 147 io" ?Oct. i9 ? Oct. 152 155 (i) i i imm. ? May/June 1(2) 93 fulviventris .... * 95 Alseonax cassini ...... 98 Chloropeta natalensis major .... 99 Parisoma plumbeum orientate 104 Dioptrornis brunneus bailundensis. _ 109 Batis m. margaritae ..... * no minulla ...... in Platysteira peltata mentalis .... I i I ? PSept. "4 i? PJuly IcJ ? Sept. 118 131 I 66", 2? Summer I $ Summer 134 i juv. i ?> July 271 i imm. 273 i$ ? July 285 i imm. July 288 IcJ, 2? ? Sept. 296 386 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS TABLE III. Birds of Grasslands, Other references Eastern Districts Central Tables Syst. Notes i 2 3 Breeding 4 5 6 5 Francolinus levaillantoides jugularis albogularis meinertzhageni . * 2 6", 3? Summer 8 Pternistes afer crainchii IV 16" ? June 10 Glareola pratincola fullebomi i? Summer 12 Cursorius temminckii .... i juv. 6" 1 8 Centropus toulou grilli .... 26 Halcyon albiventris .... i 38 Caprimulgus natalensis mpasa (?) . * i ,y Summer 39 fossii .... . 2 6", 2 $ Aug. onwards 42 Micropus caffer ansorgei 16" 64 Mirafra africana occidentalis * occidentals $ pallida . kabalii 26", 2$ Summer _ 65 angolensis angolensis . * i? 36", i $ antonii 66", 3? Summer 66 rufocinnamomea fischeri i? i 6" 69 Certhilauda albofasciata obscurata . 36", 2$ 70 Calandrella cinerea saturatior 26", 2? July 2 JUV. 72 Eremopterix verticalis damarensis . IV 73 Motacilla capensis simplicissima 74 Anthus vaalensis neumanni . - 1$ i~i Summer r? 26", I $ ~9 75 leucophrys bohndorffi 78 Macronyx ameliae .... I 6", 2 ?, 16" ? July onwards 26" I imm. $ 79 grimwoodi .... I Summer i 6* 1 68 pipiens Congo .... i (j Summer 170 bulliens 171 emirii bailundensis . 172 lais namba .... _ _ _ _ 173 rufilata ansorgei 16" 174 textrix mystica ! *! * 8 <"* i M ^ ft 1 It 111 I 'll I e"o ** r* * <^J^eJ > ^8 o 0-8, tq ^ 8 ^ S. 05 ^ ^ OH Moco and Mombolo .x.x.x.x.x.x. x Fernando Po . . .x. .y. .x. Cameroon Mt. . . .x. y.y. y.y.y Mts. of north Cameroons .x. y.y. y.y. y Ruwenzori area . .y.y.y. . . .x Mts. of Kivu and L. y.y. y.y. . . x Edward Mts. east of L. Tan- . . . x . . . x ganyika Mts. of Abyssinia . y . . . y . . . Mts. of East Africa .y.y. Y Y Y Y Mts. of S. Sudan . . . . . y . . . Mts. of north Nyasaland Y Y Y Y Mts. of south Nyasaland .y. .y.y. . x, same subspecies as on Mt. Moco. y, different subspecies from Mt. Moco. provided to show the distribution of the other seven species, all of which are confined to montane forest and are representatives of the relic montane fauna discussed by Chapin (1932 : 285) and Moreau (1952 : 888). The Angola population is subspecifically distinct in only four of these species in spite of the distance separating it from the nearest mountains in the Cameroons and Nyasaland. The distribution of the birds found in the escarpment forest at Gabela and in the forest at Chingoroi emphasizes their northern association. Some, such as the shrikes Chlorophoneus multicolor and Lanius mackinnoni and the wattle-eye Dyaph- orophyia concreta, range from West Africa along the northern edge of the Congo forest to East Africa, though with by no means a continuous distribution. The blue- bill Spermophaga ruficapilla follows the southern edge of the forest to East Africa, while the akalat Malacocincla fulvescens and the greenbul Andropadus latirostris embrace both ranges. Two other greenbuls Chlorocichla falkensteini and Phylla- strephus fulviventris, the Red-crested Turaco, T. erythrolophus, and the bush-shrike THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 393 Telephorus viridis have limited ranges in the west from the Gulf of Guinea to their southern limits at Chingoroi. The Crimson-wing, Cryptospiza reichenovii, is found unexpectedly at Gabela rather than on Mt. Moco since it is a montane forest species in the Cameroons, Fernando Po, and in most of East Africa. The most striking feature of the distribution of the Angola forest birds is the almost completely different avifauna of Mt. Moco and Gabela, the only species obtained by ourselves and /or Heinrich in both forests being the wide-ranging Tambourine Dove, Turtur tympanistm, the bush-shrike Telephorus viridis, and Colius castonotus. This is the more surprising in view of the small percentage of Moco birds that are elsewhere restricted to mountains, others being found in the northern forests of Pungo Andongo, Ndala Tando and Canzele where most of the Gabela species are also found. The two localities are only 120 miles apart and the one-time montane forests of Mombolo were even closer to Gabela. Since the montane forest is recognized as being older than the lowland forest of the Congo Basin (Moreau 1952 : 908) it suggests that the escarpment forests may also be of a relic type but have been more subject to invasion by lowland species from the north than the more isolated montane forest. The distribution of the birds also points to interesting affinities between the forest remnants in the Leba and Humpata areas and the montane forest of Mt. Moco, and also between the montane forest and some forest patches of north-western Northern Rhodesia. Grassland Birds (Table III) The distribution of the grassland birds found in Luacano and Vouga is quite fascinating and, to me, inexplicable. Typical species of these grasslands are the three longclaws, Macronyx ameliae, M.fulleborni and M. grimwoodi, the widow Coliuspasser progne, and the Angola Lark, Mirafra angolensis, all found at both places ; the Spike-heeled Lark, Certhilauda albofasciata, though collected by us only at Vouga as been found also in the eastern districts near Vila Luso. Thollon's Chat, Myrmeco- cichla tholloni, was found at Vouga only, and Francolinus albogularis at Luacano only. All of these, when present in the area, were found within short distances of each other and would therefore be expected to have the same requirements and therefore similar ranges throughout Africa, yet each is completely different. Among the long- claws the yellow M. fulleborni is the local variant in Angola, Northern Rhodesia, and the southern Congo of the widespread M. croceus ; the redthroated M. ameliae is primarily an East African species spreading westwards through Northern Rhodesia to its western limit in central Angola ; the larger M. grimwoodi was known only from north-western Northern Rhodesia until we found it first on the road to Luacano and later as far west as Vouga. The Angola Lark is confined to the highlands of central Angola, the grasslands of the Zambesi /Congo watershed and north-western Northern Rhodesia, and the Marungu highlands of the south-eastern Belgian Congo. The Spike-heeled Lark is essentially a bird of the dry western semi-desert areas of South West Africa, and the pans and grasslands of Bechuanaland and the Transvaal. It has recently been found by Rudebeck on the Humpata plateau and in the coastal strip north of Mo9amedes. 394 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS The francolin is West African, known for a long time only from scattered localities in the hinterland from Senegal to Nigeria, but recently discovered on the Marungu highlands, eastern Belgian Congo and in the Balovale area of Northern Rhodesia. Thollon's Chat is limited to the grasslands of the western plateaux of the French Congo and Angola, between the equator and 13 S. The Long-Tailed Widow has a broken distribution, being found on the high veld grasslands of the Transvaal and neighbouring provinces, in a strip across central Africa from Lake Bangweola to central and southern Angola, and in the highland grasslands of Kenya. From our own limited observations and from those of C. W. Benson (in litt.), and C. M. N. White (1958) in Northern Rhodesia it seems that even in very limited areas the grassland species are segregated to a large extent by preference for moister or drier conditions. Thus the Spike-heeled Lark, C. albofasciata, a species with a preference for sandy soil, short grass, and poor vegetation, was found at Vouga on the higher, and therefore well-drained, parts of the undulating grasslands east of the village. It may be this factor of drainage which enables these larks to tolerate here an unusually high rainfall, over 40 inches a year, whereas in most of their range the rainfall is under 10 inches a year. Similarly, lack of good drainage on some of the flatter grasslands of the eastern districts would account for their absence there. M. angolensis seems to require rather moister conditions than the Spike-heeled Lark, being collected at Vouga on the edges of the marshy valleys and by White at Balovale on the midslopes of the Minyanya plain. At Minyanya the ecological place of the Spike-heeled Lark on the upper slopes is filled by M. africana, which was found alongside M. angolensis at Luacano, where the limited amount of grass- land in the surrounding districts and the uniform flatness of the plains may have brought the species together through lack of their preferred habitats. The francolin, longclaws and widow-bird were associated with the dampest areas of the grasslands. The widows were in flocks along the ditches at Luacano, though in north-eastern Northern Rhodesia Benson found them associated with M. africana on the drier grasslands. All the longclaws were collected in similar conditions in very rank grass and in the marshy valleys. They seemed in Angola to be competitive species since all had been feeding on grasshoppers and beetles : Benson, however, found that the yellow species fulleborni and croceus in Rhodesia prefer drier parts than do the red-throated species, M. ameliae and M. grimwoodi. It may be that the widespread yellow species are more adaptive to differing conditions than the red- throated species, especially M. grimwoodi. Some of the francolins were found to have been eating frogs and fish, the latter possibly among those that had been left stranded with the drying up of the summer floods. It is most improbable that fish or frogs form an essential part of the diet of F. albogularis (others collected by Serle in Nigeria had been feeding on cassava), but high rainfall does seem to be a primary requirement for the species since the range lies entirely in the area between the 4o-inch and 6o-inch isohyets. It is easy to see therefore, why all the species of the wetter parts of the grassland should be absent from the dry grasslands of south-western Africa, but not apparent why each should have such differing ranges within the wetter areas, the broken THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 395 distribution of Coliuspasser progne being particularly surprising. From our observations of Thollon's Chat it would seem unlikely to be found except where there are occasional small trees or bushes on which to perch, these perches being used as observation posts to which the birds returned time and again after foraging. Lack of suitable perches would account for their absence on the Luacano plains, and if all the grasslands of the eastern districts are similar to those we crossed, this may be the limiting factor in distribution of the species eastwards. Among other families classed in Table III are many cisticolas and members of the Finch family that are birds of grasses and reeds rather than open grassland. Most of these are wide-spread but many of them have local and very distinctive races in central and western Angola. Birds of the Western Thornbush, Semi-desert, Rocks and Tangled Thickets (Table IV) In contrast to the complexities of distribution among the grassland birds of the central and western districts the ranges of all western species follow regular and comprehensible lines governed by the topography of south-western Africa and the limits of the thorn veld. Thus, with the exception of Euplectes aurea which was possibly introduced from Sao Tome, all the species collected in the semi-desert areas are found in similar areas in South West Africa, though in many cases are racially distinct. Typical thornveld species such as the Pied Barbet, Tricholaema leucomelas, and the gregarious larks such as Eremopterix verticalis, have a wider distribution through the Kalahari and have been found, or can be expected, in Angola as far north as Luanda (see Serle, 1955), while species more typical of the Vornamib scrub such as Schlegel's Chat, Karrucincla schlegeli, are unlikely to be found north of Lobito. Similarly, among the rocks of the Chela escarpment are such species as the Pale- winged Starling, Onychognathus nabouroup, [the rock-jumper, Achaetops Pycno- pygius, and the Chat-flycatcher, Lanioturdus torquatus, all of which have limited ranges in the highlands of south-western Africa. The range of the indigenous Xenocopsychus ansorgei is even more limited, being apparently restricted to the rocks of the Chela escarpment and in the neighbourhood of Mt. Moco. The overall pattern is of the predominantly South West African avifauna petering out northwards along the coast and the edge of the escarpment. AVIFAUNAL ZONES AND SPECIATION Avifaunal Zones The distribution of birds found in the different types of vegetation has been discussed at some length because it seems that an understanding of this factor is primarily important in assessing the relationship between many forms found in Angola. Repeatedly it will be found that birds of many different families have similar specific or sub-specific ranges which lie along the same boundaries and coin- cide with the limits of one or another type of vegetation. Thus western Angola can be divided into three avifaunal zones (see Map) : Zone i. Above the escarpment and coinciding with the limits of Brachystegia woodland from the Sa da Bandeira area north at least to Malanje. This I shall call 396 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS the Brachystegia Zone and in it East and Central African species and subspecies predominate. Zone 2. The Acacia Zone, including the desert and semi-desert and poorer parts of the coastal plain, and the thorn country on top of the escarpment south and west of the Brachystegia Zone. In this South West African species and subspecies predominate. Ben Mocamedes ESCARPMENT Vila Salarar Miles Zone 3. The north-western coastal lowlands where a richer and more varied vegetation is found than in Zone 2, and the escarpment area including the escarp- ment forests. This I shall call the Escarpment Zone. Little collecting was done in it except in the forests at Chingoroi and Gabela and therefore the distribution of birds, other than forest species, has not been discussed ; it would seem however to contain a number of interesting endemic forms about which much remains to be learnt. These include the two newly-discovered species Prionops gabela and Muscicapa gabela, both found originally in secondary forest and the Prionops since found in thickets below the escarpment : also the warbler, Macrosphenus pulitzeri, THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 397 known only from the type specimen, and the forest greenbuls, Chlorocichla falken- steini and Phyllastrephus fulviventris. Among endemic forms are three which have been discussed in the Systematic Notes, Malaconotus hypopynhus blanchoti, Camar- optera brevicaudata harterti and Tchitrea melampyra bannermani. The Escarpment Zone may even be found to extend in pockets as far south as Vila Arriaga and Capangombe as has been suggested in the Systematic Notes in the discussion on the distribution of the two races of Lamprocolius mevesii. The interest of these zones lies in the fact that the zonal limitations are applicable to some birds that are not primarily associated with the typical vegetation ; for example Cercomela familiaris is a bird of rocks and kopjes rather than woodland or thornbush, but the two races falkensteini and angolensis are confined respectively to the Brachystegia and Acacia Zones. Similarly, the Black Tit, Parus leucomelas insignis, is found only outside woodlands in clearings or thickets but its true range is limited to Brachystegia Zone with an atypical population in the Escarpment Zone but not probably in the forest. It should, in fact, be emphasized that the zonal limits of the Escarpment Zone probably apply less to typical forest birds than to others, for several forest species are found in patches of riverine forest and in montane forest, within the Brachystegia Zone. On the Map the zig-zag line dividing Zones i and 2 is intended to indicate that the two zones interlock where patches of Acacia and Brachystegia alternate. Some sub-divisions of Zones i and 2 have been shown, which were indicated in the discus- sion on distribution. No attempt has been made to define the zonal limits in the northern and south-eastern parts of the country from which little or no information on the birds is available. Speciation The rapid changes of vegetation and of altitude in the west of the country which are responsible for forming the zonal boundaries and the pockets of forest, allow closely related forms associated with specialized habitats to live in proximity without interbreeding. These forms can either be closely allied species or subspecies. A typical example of representation by species is found in the Batis flycatchers, in which B. molitor is the representative in the Brachystegia Zone, B. print in the Acacia Zone, B. minulla in the Escarpment Zone and possibly also in riverine forest in the north, and B. margaritae in montane forest, while the two races of shrike, Lanius collaris capelli and L. c. subcoronatus provide examples of subspecies in the Brachy- stegia and Acacia Zones which do not apparently intergrade. Between good species like those of the genus Batis and subspecies like the shrikes there are other couplets or triplets of zonal representatives or forms ecologically separated whose relationship cannot confidently be defined. It would seem that through segregation they are in the process of developing into good species and are at varying stages in the process. In recent studies on various African birds the problem has been discussed of forms which behave as good species in one part of their range and as subspecies elsewhere, by authors such as Chapin, Moreau, Benson, White and Stuart Irwin. It will be found that for most of the birds considered western Angola is a critical THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS area in which speciation seems more advanced than in other parts of Africa, and it is hoped to develop this conception more fully in a later paper. Meanwhile the following list attempts to place in order some Angolan pairs or groups according to their degree of speciation, with the examples heading the list of forms that are no more than well-differentiated subspecies, grading through those of more doubtful relationship to some that behave as true species at the end of the list. Some of these pairs are zonal representatives, for the others there is some other ecological factor causing segregation which is indicated if possible. Reference is also made to papers in which the relationship has been discussed. The conventions of trinomial nomenclature compel one to make a dividing line between pairs considered subspecies and pairs considered species, although for forms in the centre of the list there are arguments for and against either treatment ; it seems impossible to attain consistency since each problem is different and is aggravated by lack of data on such characters as voice, behaviour, nest and eggs. Far more also will be known about the relationship between some of the pairs in the Brachystegia and Acacia Zones when adequate collecting has been done along their boundary in the south and south-eastern districts of Angola. Only in one example quoted, the Spur-fowl, Pternistes afer, does it seem possible that the process of speciation has been reversed and that, instead of forms evolving differences through isolation in different habitats, two very dissimilar forms afer and crainchii, which once may have appeared good species, have come together and hybridized and must now be considered conspecific. This suggestion of hybridiza- tion is put forward tentatively since this is the only case in western Angola in which intermediate populations are found, not apparently associated with vegetational changes, linking two highly differentiated forms of the thorn and Brachystegia country. Furthermore, these intermediate populations (to which the names benguel- lensis Bocage, chio Meise, and cambatelae Meise, have been given) show a degree of individual variation more commonly associated with hybrid populations than with populations of an established species. EXAMPLES OF STAGES OF SPECIATION Erythropygia leucophrys manda (bushes in Zone i) Erythropygia 1. ovambensis (bushes in Zone 2) Prodotiscus i. insignis . (Zone 3, forest ?) Prodotiscus i. lathburyi (Zone i, highlands) Prodotiscus i. zambesiae (elsewhere) Zosterops senegalensis quanzae (highland in Zone i) Zosterops s. under ssoni (Zone 2) Zosterops s. heinrichi (Zone 3) Subspecies showing intergradation. P. i. lathburyi a subspecific link between, insignis and zambesiae. Subspecies showing intergradation. See Moreau, 1957, an d Systematic Notes. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 399 Pternistes afer afer . (Zone 2) Pternistes afer crainchi (Zone i) Prinia flavicans bihe . .- (bushes on Vouga grasslands, Zone i) Prinia f. ansorgei (thornbush, Zone 2) Mirafra africana occidentalis . (grasslands of Zones 2 and 3) Mirafra a. kabalii (grasslands of eastern Zone i) Nilaus afer affinis .... (highland Brachystegia, Zone i) Nilaus a. brubru (thornbush, Zone 2) Laniarius luhderi amboinensis (escarpment forest ?, Gabela, Zone 3) Laniarius I. brauni (escarpment forest, Quicolungo, Zone 3) Eremomela icteropygialis polioxantha (woodland in Zone i) Eremomela i. puellula . (thornbush in Zone 2) Eremomela salvadorii (woodland Zone i) Lanius collaris capelli .... (bushes in Zones i and 3) Lanius c. subcoronatus (bushes in Zone 2) Malaconotus hypopyrrhus monteiri . (Zone 3) Malaconotus h. interpositus (bushes, Zone i) Lamprotornis mevesii mevesii (mopane belt, Zone 2) Lamprotornis m. benguelensis (southern tip of Zone 3) Cercomela familiaris falkensteini (rocky hills in Zone i) Cercomela f. angolensis (kopjes in Zone 2) Camaroptera brevicaudata harterti . (forest undergrowth, central Zone 3) Camaroptera b. sharpei (undergrowth, Zone 2 and southern Zone 3) Camaroptera brevicaudata near sharpei (undergrowth, Zone i) Camaroptera b. tincta (undergrowth, northern Zone 2) ZOOL. 6, 7. Well-marked forms linked through hybrid populations. See above, and Bowen, 1930. Probably well-defined subspecies, isolated from each other. See Irwin, 1959. Well-differentiated forms probably sub- species : widely separated geographically. See Systematic Notes. Well-differentiated forms often considered species but with some intergrades recorded. Exceptionally well-differentiated forms about which little is known. The first two well-defined subspecies ; salva- dorii probably a closely related species over- lapping in range. See Systematic Notes. Well-defined forms with close ranges but no recorded intergradation. Well-defined forms with close ranges but no recorded integradation. See Systematic Notes. Well-defined forms living in very close proximity ; no apparent intergradation. See Systematic Notes. Well-defined forms living in very close proximity; no known intergradation. harterti possibly a species. Notes. See Systematic 24 400 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Tchiirea melampyra bannermani (forest, Zone 3) Tchitrea viridis plumbeiceps Zones i and 2, also 3 outside forests) Passer griseus griseus .... (habitations widespread) Passer diffusus diffusus (away from habitations, Zones 2 and 3) Lagonosticta landanae .... (Zone 3 and western Zone i) Lagonosticta rubricata congica (eastern Zone i) Parus niger carpi .... (Zone 3 outside desert and semi-desert) Parus leucomelas insignis (typical) (Edges of woodland, Zone i) Parus leucomelas insignis (atypical) (Zone 3) Parus griseiventris .... (woodlands in Zone i) Parus afer benguelae (semi-desert, Zone 2) Laniarius aethiopicus major . . (undergrowth in woodlands, Zone i) Laniarius bicolor guttatus and sticturus (undergrowth, reeds etc. in Zones 2 and 3) Pycnonotus nigricans .... (Zone 2) Pycnonotus barbatus (Zone i) Elminia albicauda .... (woodlands and forest, Zone i) Elminia longicauda loandae (forests, Zone 3) Prionops gabela ..... (forest and thickets, Zone 3) Prionops retzii (elsewhere) Turdus libonyanus verreauxi . (woodland and bush) Turdus pelios bocagei (forests of west) Turdus olivaceus stormsi (forests of east) Monticola angolensis .... (woodland and bushes, Zone i) Monticola brevipes (rocky terrain, Zone 2) Possibly hybrid populations originally, now re-established as sympatric species. See Systematic Notes. Well-differentiated forms with possibly over- lapping ranges. See White & Moreau, 1958, and Benson, 1956. Well-differentiated forms with close ranges in Lower Congo : considered conspecific by Chapin. carpi a well-differentiated form living in close proximity to P. leucomelas. See Hall, 1959. griseiventris, probably good species known to overlap elsewhere with P. afer, but possibly isolated from it in Angola by the Mopane belt. See Hall and Traylor, 1959, and Irwin, 1959- Probably good species known to overlap else- where through geological segregation, living in close proximity in Angola. See Hall, 1954- Apparently good species but with possible hybridization in western Angola. See Irwin, 1958. Species living in close proximity on the zonal boundary. Closely allied species, but with little data on their relationship. T. libonyanus a good species ecologically segre- gated from the others which are possibly conspecific. Good species, closely related, representative of each other in their respective Zones. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 401 BREEDING SEASONS IN ANGOLA General Discussion From a collection such as this, taken only during the months of August and September and with a limited number of specimens of each species, generalizations on the regular breeding seasons must necessarily be limited ; nevertheless some useful data are provided by the presence of young birds and by the condition of the gonads and plumage, as to the probable breeding months of individual birds. In the Tables of Specimens Collected an estimate of the approximate breeding month has been made in all cases in which there appears to be sufficient evidence to warrant it. Nest with eggs, and young birds not fully grown, provide the only incontrovertible evidence these young birds are listed in the Tables as " juvenile ;" young birds, fully grown, are classified as " immature." Some of these immature birds, still with considerable amounts of young plumage, also give some certainty within one or two months, especially when they are supported by adults in post-breeding moult. The estimated breeding month of these, as with juveniles and eggs, is put without a query in the Tables. The date of breeding of older, but still not fully adult pas- serines, is far more uncertain and has been indicated vaguely as ? May /June or ? June /July to show that it probably took place at the ends of the rains (see Table of Rainfall) or the beginning of the dry season rather than in the early or middle part of the rainy season. Specimens in which the skull is not fully ossified are also classed as " immature ", but my knowledge of the length of time that full ossification takes in different families is too limited to allow me to make an estimate of breeding season on this evidence alone. Among sunbirds in particular a high proportion of apparently adult birds were noted as having ossification incomplete. The condition of gonads and plumage has also been considered in conjunction, though it is appreciated that both can be misleading in birds of the tropics. Those specimens in worn plumage with active gonads have been considered as possible breeders in the month or two months following collection, i.e. in the early rains ; those in very fresh plumage, or post-breeding moult, showing no breeding activity, as possible breeders two months prior to collection, and therefore during the dry season or winter ; those with non-active gonads, in indeterminate plumage, non- breeding dress, or moulting from non-breeding dress have been classified as " sum- mer " breeders on the assumption that they had not bred recently and are unlikely to do so before November. On these estimates it is possible to make some cautious contributions to the tables of breeding activity in Africa published by Moreau (1950) . It would be expected that Angola would fall into the regular pattern of the countries of southern Africa which have summer rainfall. In these countries, which include the Rhodesias, Nyasaland, and northern parts of the Union, the peak of breeding activity is found in September to December, that is in the early months of the rains and the two months preceding the rains, while in the dry winter months of May to August there is very little breeding activity except amongst birds of prey. In South West Africa, also an area of summer, but erratic, rains Macdonald (1957 : 30) found a progressive 402 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS retardation of the breeding season from south to north with evidence that birds of Damaraland and the Kaokoveld have a peak activity from December to March. It was therefore unexpected to find evidence of considerable breeding activity in Angola in the dry winter months of June to August. Young birds of over fifty species were collected in August and September. These were found in all the main vegetational regions and were of widely assorted families. In addition nest and eggs were found in August of the Stone Chat, Saxicola torquata, and the martin Petro- chelidon rufigula. Some rainfall figures relevant to the discussion are included in the Tables on p. 403 : Non-passerines Comparatively few non-passerines were collected and these for the most part indicated the expected summer breeding seasons in all species in the eastern and central districts whether woodland, forest, or grassland, except for a juvenile Cursorius temminckii collected west of Munhango on i2th August, and a juvenile Colius castanotus on Mt. Moco. In the western districts, in the semi-desert areas, the courser, hornbill and coly were in post-breeding moult : in the grasslands and evergreen thickets juveniles were obtained of Pternistes afer, and Francolinus leva- illantoides. The passerines are best discussed by ecological groups and districts according to the Tables. Passerines of the Savannah Woodlands (Table I) The flycatchers, thrushes and warblers generally indicate an August /September or a " summer " breeding season, sometimes slightly in advance of the corresponding families in Rhodesia, but similar to them in relation to the onset of the rains which start a little earlier in Angola. The shrikes and finch family appear irregular with some dry-season activity. The sunbirds, as in Northern Rhodesia and the southern Congo, seem frequently dry-season breeders. Passerines of the Forests (Table II) In the eastern riverine forests a number of indications of early winter or late summer breeding is provided by immature but nearly fully adult specimens, with rare indications of spring or early summer activity. On Mt. Moco, on the contrary, most birds had active gonads in August, and the Striped Swallow, Hirundo abyssinica, and Seicercus laurae were observed apparently engaged in nesting activities. The only winter breeder was apparently the weaver Hyphanturgus ocularis. In the western forests the indications were of a peak season in June /July, very few species collected not being represented by either a juvenile or immature specimen; many of the latter among the thrushes and finches were still largely in first plumage although the shrikes and bulbuls were apparently older, the indications of immaturity being in the bill colour, skull ossification, or in a few remaining immature feathers. It would therefore seem that in these patches of forest the birds of the trees and forest canopy have a breeding season slightly in advance of the birds of the under- THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Some Rainfall Figures Relevant to the Discussion Measured in Inches, Less Than One Tenth of an Inch Shown as Trace " tr " Rainfall in 1957 (Localities on the Expedition Route) Locality Years J. F. M. A. M. J. J. A. S. 0. N. D. Year's total Texeira de Sousa . 1957 10-2 5-9 9-6 0-3 ? ? 2-1 4-8 3-6 8-3 44-8 plus Nova Lisboa 1957 6-5 5-9 6-7 2-8 tr. tr. tr. 2-0 2-6 9-2 13-1 48-8 Sa da Bandeira 1957 3-5 4-1 12-6 5-2 0-4 8-7 6-2 8-0 48-7 Lobito 1957 tr. o 9 1 1 6 7-4 tr. 0-3 0-4 2-8 1-4 24-8 Gabela* 1957 1-5 5-3 n-8 5-5 0-4 2-9 4-8 5-4 37'6 Quimberima . 1957 First spring rains fell 27th September. Mt. Moco No figures available : first spring rains fell very locally on 24th August. Monthly Averages in Eastern Angola and North-western Northern Rhodesia Locality Years J. F. M. A. M J. J. A. S. O. N. D. Yearly average Texeira de Sousa, 3,609 feet Mwinilunga, 4,447 feet 1942-53 1921-55 8-8 8-6 9-3 5-2 0-5 tr. 0-3 0-7 3-1 7-9 8-4 8-9 8-0 8-1 3-3 0-3 o-i tr. tr. 0-6 3-7 7-5 9-5 52-8 50-0 Balovale, 3,577 feet 1920-55 8-5 6-9 5-8 1-2 tr. tr. 0-3 2-3 4-4 8-9 38-3 Monthly Averages in Central Angola Locality Years J. F. M. A. M. J. J. A. S. O. N. D. Yearly average Nova Lisboa, 5,577 1940-54 8-7 7-8 9-8 5-7 0-4 tr. tr. 0-6 5-5 9-6 8-9 57-o feet Ganda, 4,790 feet, 1924-52 7-0 7-1 10 -9 6-1 5-4 tr. i -i 6-0 9-3 8-7 61-8 55 m. N.E. of Chingoroi Gabela,* 3,586 feet 1947-55 2-9 4-6 6-8 10 -2 1-4 tr. 0-3 2-0 5-0 3-4 36-6 Quimberima, 3,700 1954-57 4-4 i-i 7-7 8-5 0-8 1-4 5-5 io-o 5-8 45-2 feet (Aug.) Monthly Averages in South-western Angola and South West Africa Locality Coastal Years J. F. M. A. M. J. J. A. S. 0. N. D. Yearly average Lobito Mossamedes Walvis Bay . 1932-53 1930-53 1916-50 0-8 1-5 4-7 2-1 o-i tr. o-i 1-2 i o 2-4 0-3 0-4 0-7 0-5 tr. tr. tr. tr. tr. tr o i o i tr. o 2 0-3 o-i o-i tr. tr. o i tr. tr. tr. tr. 13-9 2-1 0-9 Inland Sa da Bandeira 5,860 feet Ondonga,3,593 feet 1932-53 1902-50 5-9 6-2 6-6 3-7 o-i tr. 0-2 2-8 4-4 6-3 4-3 4-8 3-7 1-4 o-i tr. tr. o-i 0-5 1-8 3-6 36-2 20-3 * Rainfall figures for Gabela refer to the town on top of the escarpment, not to the forest on the face where the precipitation is higher. Compiled from the Meteorological Office Tables of Temperature, Relative Humidity and Precipitation for the World, Pt. IV, 1958, with additional figures supplied by the Meteorological Office, Harrow, and by Mr. O. Kroel for Quimberima. 404 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS growth and bushes. Unfortunately, Heinrich (1958) has few records of breeding activity for the same species collected at Gabela ; such as there are conform with the general pattern of our collections, with young birds of Spermophaga ruficapilla collected at the end of July, and other species noted as having inactive gonads in August. On the other hand, he records breeding activity in several of the same species collected in similar forest further to the north at Canzele in March /April, most notably young birds then being found of Dyaphorophyia concreta, Lanius mackinnoni and Camaroptera brevicaudata. Passerines of Grasslands, Reeds and Open Country (Table III] In the east and centre the larks, chats, pipits and long-claws of the grasslands at Luacano, Lake Dilolo, Vouga and along the road mostly appear to be summer breeders. The exceptions were the Red-Capped Lark, Calendretta cinerea, of which juveniles were collected early in August, and Saxicola torquata which was nesting at Vouga. The Cisticolas, without exception, were in non-breeding dress or in moult to breeding dress, showing themselves summer breeders, as Lynes found. Nests and eggs of Petrochelidon rufigula were found on the loth August in contrast to the record from the Belgian Congo of breeding in April, and a specimen of Hirundo sene- galensis in moult also suggests dry season breeding. The most surprising records how- ever concern the bishops, widows and waxbills, among which are many young birds and adults moulting into non-breeding dress indicating a breeding season continuing at least into June. The same species in the southern Congo and Northern Rhodesia are mostly recorded as summer breeders finishing in May at the latest. It is difficult to account for the contrasts and the similarities shown by different families in comparison with what is found in their relatives in adjacent countries, for example why the cisticolas should be normal and the bishops abnormal. The grass burning must inevitably affect the grassland birds ; this had mostly been completed in the east but was being carried out in August in the central highlands. It may be significant that it is the seed-eaters which had apparently finished breeding before the fires while the insect-eaters bred after the fires. In the western grasslands among the few specimens collected was a young Mirafra africana in September, as well as the young Francolin already referred to. On Mt. Moco the waxbills and cisticolas living in the rank grass of the mountain gullies also had contrasting breeding seasons, the waxbills breeding in winter (a number of young birds being obtained in August) and the cisticolas in summer. Passerines of the Western Thornbush, Thickets and Semi-desert, (Table IV) The number of specimens collected in the thornbush and tangled thickets on top of the escarpment in the south-west and on the road north from Lobito was limited, but it is apparent that the breeding season is very varied among the different species. In addition to the numbers of young spur-fowl seen and collected, young of Laniarius bicolor, Fringittaria tahapisi, Apalis flavida and Alseonax cinereus were collected in August, but other species, even in the same families, gave evidence of being summer breeders. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 405 The pattern in the semi-desert of the coastal plain is very different, with a high proportion of species in very fresh plumage or post-breeding moult. In arid country such as this the plumage fades and abrades very quickly so that when a specimen is found to have very fresh colours and unabraded feathers it cannot long have completed moult and the breeding season can therefore be assessed with greater assurance than with birds living in more sheltered conditions. These birds collected in September seem likely to have bred in June /July, and are in comparable plumage to those collected in the Kaokoveld in May. This three months' difference was unexpected, since the dry coastal belt of Angola seems a natural continuation of the Namib and Vornamib of South West Africa with no natural barriers between (the break made by the Cunene can hardly have significance). It is too long also to consider as a further logical step in the retardation of the breeding season from south to north as noted by Macdonald. However, the explanation may possibly lie in the rainfall figures for 1957, which show that in March and April there were exceptionally heavy rains at Lobito, 1 1 -6 inches and 7-4 inches against an average over twenty- two years of 4-7 inches and 2-1 inches for the same months. The figures at Benguela are not so high, 7-6 inches in March and 3-3 inches in April, but are still exceptional for that coast. It seems quite probable that there would be heavy casualities among nests, eggs and young birds under these conditions, and that second broods were therefore reared. Data on breeding seasons in other years on this coast is badly needed before the problem is really understood ; a little is provided by specimens in the British Museum collected by Ansorge in 1905 among which are some of Erythropygia paena, Plocepasser mahali and Passer motitensis in post-breeding moult at the end of July, suggesting an earlier breeding season in 1905 than in 1957. NEW RECORDS FOR ANGOLA AND EXTENSIONS OF RANGES Species not Hitherto recorded from Angola Francolinus albogularis meinertzhageni Luacano. Francolinus sephaena zambesiae Leba. Mirafra passerina Cahinde. Macronyx grimwoodi Luacano area, Vouga. Extension of Ranges in Angola Poicephalus meyeri matschei Luacano (western extension). Caprimulgus natalensis mpasa Lake Dilolo (western ext.). Andropadus virens zombensis Luau R., Luacano (western ext.). Alseonax c. cinereus 40 m. S. of Mumbondo (southern ext.). Cossypha h. heuglini Luacano (western ext.). Erythropygia leucosticta reichenowi Chingoroi (southern ext.). Eremomela icteropygialis polioxantha Vouga, and 100 m. West of Vila Luso (western ext.). Prinia sub/lava ovampensis Quipungo, Leba (northern ext.). Campephaga q. quiscalina Chingoroi, (southern ext.). 406 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Prionops gabela 40 m. S. of Mumbondo (second known locality). Chlorophoneus nigrifrons manningi Luau R., (western ext.). Cinnyris s. superbus Chingoroi (southern ext.). Cyanomitm verticalis bannermani Luacano (western ext.). Euplectes orix subsp? Lake Dilolo (northern ext.). Lagonosticta rubricata congica Luau R., (western ext.). Estrilda astrilde niediecki Vouga (northern ext.). Hypochaera nigenima Chingoroi (southern ext.). New Races It was found necessary to describe three new races, mainly on specimens from this collection. These were published in short papers as follows : Prodotiscus insignis lathburyi Hall, Bull. B.O.C. 78, 1958, p. 152 : Mt. Moco. Mirafra angolensis antonii Hall, loc. cit. p. 153 : 5 m. east of Luacano. Parus afer benguelae Hall & Traylor, Bull. B.O.C. 79, 1959, p. 12 m. SE. of Benguela. Sight Records Among species of interest seen but not collected were the Palm-nut Vulture, Gypohierax angolensis, common along the road between Lobito, Novo Redondo and the foot of the escarpment at Gabela. Hartlaub's Duck, Pteronetta hartlaubii, a pair on the Luau River ; the Bustard Eupodotis senegalensis mackenziei, at Vouga ; the Rufous-Bellied Heron, Erythrocnus rufiventer, about twenty-five of which spent the day in high reeds near a river at Lake Dilolo, flighting down to the main lake half a mile away at sunset. A few were also seen on a small lagoon on the Humpata plateau near Jau. The Giant Blue Turaco, Corytheola cristata, was seen on the Luau River and at Gabela, both localities representing the southernmost limits of its range as known at present, the Luau River being an extension of the range shown by Moreau (1958 : map 4, p. 72). Three other species seen at the southern limits of their ranges are Monteiro's Twinspot, Clytospiza monteira, at Gabela, the Shining Blue Kingfisher, Alcedo quadribachys, on the Luau River, and Boehm's Spinetail, Neafrapus bohmi, 80 m. south of Luanda. SYSTEMATIC AND FIELD NOTES ON SELECTED SPECIES In working on the systematics and taxonomy of the Angola collection the standard works on the birds of the neighbouring countries have all been consulted, Chapin (1932-1954) for the Belgian Congo, Benson & White (1957) for Northern Rhodesia, Hoesch & Niethammer (1940) and Macdonald (1957) for South West Africa, Roberts (1940) for South Africa as well as the manuscript of Mackworth-Praed & Grant's " Birds of the Southern Third of Africa." It is inevitable that all these authors do not always agree on systematic questions, especially the vexed one of " lumping " in genera and species. I have followed no THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 407 author consistently, using my own preference where there are differences. On the generic and specific levels I am conscious that I have not always been able to study the problems as deeply as is necessary to form a considered opinion ; I have perhaps been influenced by my mistrust of over-enthusiastic lumping, which I feel has sometimes been proposed to evade the necessity of making a decision. It is not, to me, necessarily sufficient reason for uniting two genera, that one species may show characters of both. The subspecific status of all forms listed has been studied on the evidence of speci- mens in the British Museum, supplemented in many cases by information and loans from other Museums, most notably from Mr. M. A. Traylor and the Chicago Natural History Museum, where a large part of the Angola collection from the American Museum of Natural History is temporarily housed. The Systematic Notes include any revisions that appear necessary of the accepted range of status of the various forms, and some notes on moults and plumage of the less well-known species. Most of the field-notes on sunbirds were supplied by John Williams, and the majority on all other birds by Tony Archer and General Lathbury. Williams was also responsi- ble for most of the detailed notes on food, derived from the contents of stomachs and crops ; but in the case of the turacos, some francolins, and Xenocopsychus, detailed analysis of the seeds and insects was made in the Botanical and Entomological Departments of the Museum. Since the collection was made papers have been published on collections from Angola by Heinrich (1958), and Meise (1958). These contain field or systematic notes on many of the species we obtained and in selecting species in the following notes I have endeavoured to avoid duplicating information and to concentrate on those in which some contribution can be made to present knowledge of the bird in the hand or in the field. Data on all other specimens are, however, readily available in the British Museum should they be required. With so many standard works now available on African birds it has not seemed necessary to provide other reference than the author's name and date to the forms discussed unless they have been recently described or the names newly brought into use. Type localities and measurements are included only when relevant to the discussion. 2. Francolinus swierstrai (Roberts), Ann. Trans. Mus. 13, p. 72, 1929 : Mombolo, Cuanza Sul. (syn. Francolinus cruzi Themido 1937 : Hanha, Benguela dist.). Specimens i <$, i imm. <. Mt. Moco 7,000 feet, i7~22nd Aug. Bill and legs orange red ; iris hazel. Wing ad. 184, imm. moult ; bill 32, 30 ; tail 103, 84 mm. Field Notes This rare francolin was found in or near pockets of forest on the upper slopes. One was found pecking about among fallen leaves in the undergrowth, the other 408 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS was flushed from the bank of a stream covered in thick bracken. It uttered a shrill harsh cry, not unlike that of Francolinus jacksoni O-Grant, and perched in a low tree. The crops contained seeds of two species of Leguminosae, one of which was a Dolichos or Neomutanenia, the other unidentified. Systematic Notes One of these specimens is an adult male with a long spur ; it is in worn plumage with some fresh feathers coming in on the head. The second is not quite fully adult, having a short blunt spur and some rusty edges to the wing-coverts, as described by Roberts for the type, which was also a young male. The young bird is mostly in fresh plumage with the primaries in moult and some fresh feathers coming in on the head. The gonads of both specimens were slightly enlarged. 3. Francolinus sephaena zambesiae Praed 1920 : Mesanague, Zambesi, P.E.A. Specimen i $ Leba, 2nd Sept. Systematic Notes A soil-stained specimen but matching well in markings with the type and others of this race, of which F. s. thompsoni (Roberts) is considered a synonym. 4. Francolinus finschi Bocage 1881 : Caconda, Huila dist. Specimens i cJ i $ Mt. Moco 7,000 feet, i8-2ist Aug. Field Notes These birds were collected and others were seen, often in pairs, in the Brachystegia woodlands on the upper slopes of the mountain, and on the bare slopes above the tree line. The female was one of a pair pecking about in burnt grass and leaves. The male had been feeding on buprestid beetles, larvae and beans. Systematic Notes These specimens have been compared with a male and female from Brazzaville, lent by the Museum d'Histoire Naturelle, Paris, and with a male in the British Museum from the watershed of the rivers Cuanza and Luando. The three latter are very similar in general appearance and in the extent of rufous in the wings. The Moco birds in comparison have less rufous in the wings, with none on the outer web of the first four primaries, and the rufous on the inner web not reaching to within 20 mm. of the tip on the first five primaries. It is possible that these differences may represent geographical variation. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 409 7. Francolinus albogularis (?) meinertzhageni White, Bull. B.O.C. 65, 1944 p. 7 : Kumanu Plain, west Balovale, N. Rhodesia. Specimens 2 <$ 3 $ Plains near Luacano, 3ist July-ist Aug. Field Notes Three of these White-throated Francolins were shot out of a covey of seven on an open grassy plain with no vegetation over a foot high. The others were a pair on a similar plain ; the female did not run or fly after the male had been shot. They had been feeding variously on beetles, grass seeds, small fishes and a small frog. Systematic Notes All are in worn plumage and stained with burnt grass. Even allowing for this the males appear to be darker, less rufous and more patterned with grey than the type of meinertzhageni. They show, in fact, much the same characters that distinguish the Luacano race of Mirafra angolensis, M. a. antonii, and it is possible that when better series are available from Angola and Rhodesia the Luacano francolins may also be found to represent a new race. 21. Tauraco corythaix schalowi (Reichenow) 1891. Specimens 2 $ Mt. Moco 7,000 feet, i6th Aug.: i $ Leba, 2nd Sept.: i $ Chingoroi, gth Sept. Field Notes It seems worth recording that at Chingoroi, where both T. c. schalowi and the Red-crested Turaco, T. erythrolophus, were found together, the crop of each was found to contain quite different types of seeds, that of T. c. schalowi being full of small black seeds, probably Molvaceae and possibly a species of Sida, and some unidentified flattened seeds ; that of T. erythrolophus seeds, possibly of the family Rubiaceae. 22. Tauraco erythrolophus (Vieillot) 1819. Specimen i Chingoroi, 8th Sept. Field Notes This specimen was collected in not very thick forest bordering the stream. Others were seen later in the escarpment forest at Gabela. See under T. c. schalowi for crop contents. 410 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Systematic Notes Moreau (1958 : 104) includes Huila in the range of T. erythrolophus : he has asked me to make a correction to this since the specimen in the collection at Pittsburgh on which this was based is actually labelled " dist. Huila " and was collected the same day as another labelled " Chingoroi." Chingoroi lies just within the boundary of the Huila district : the Turaco has not otherwise been recorded further south than Chingoroi, nor does there seem suitable habitat for it, and that specimen must therefore be presumed to have come from Chingoroi. 33. Otus scops senegalensis (Swainson) 1837 : Gambia. Specimen i (J 20 m. NE. of Novo Redondo, I4th Sept. Wing 125 mm. Systematic Notes This is a very small Scops Owl, very sandy in tone with fine vermiculations. In the British Museum it can be matched for colour with a specimen from Cape Town and several from West Africa. It seems very doubtful therefore if there are grounds for recognizing races on colour in a species with so much individual variation and I follow Mackworth-Praed and Grant (1952 : 650) in considering the majority of African Scops Owls as 0. s. senegalensis. This sandy specimen refutes the suggestion that birds of western Angola can be recognized as a distinct grey race, 0. s. hendersoni (Cassin), the type of which was obtained at sea off Novo Redondo. 36. Ciccaba woodfordii bohndorffi (Sharpe) 1884 : Semmio, Niam-Niam country, Belgian Congo. Specimen i Luacano, 2nd Aug. (b) i 2 i imm. $ 12 m. S. of Gabela, ly-iSth Sept. Systematic Notes The distincti veness of the north Angola population and variation within C. n. intensa have been discussed (Hall 1958 : 154). 138. (a) Erythropygia leucophrys munda (Cabanis) 1880 : Malanje. (b) Erythropygia leucophrys ovamboensis Neumann 1920 : Ombongo,Ovam- poland. Specimens (a) i $ Quipungo, 29th Aug. (b) i $ E. of Cahinde, 6th Sept.: i $ 60 m. N. of Lobito on Novo Redondo road, I3th Sept. Systematic Notes It would seem that munda and ovamboensis are ecological races, munda being found in the woodlands and ovamboensis in the thorn country. Both the Quipungo and Cahinde birds were taken in areas where the two types of country meet, the first in a bush between a pocket of woodland and a pocket of thorn scrub, the second near the foot of the escarpment where stunted Brachystegia and mopane gives way to the semi-desert. This latter specimen has also some intermediate characters being more heavily streaked than is usual for ovamboensis, but is considerably paler above and less washed with rufous below than munda. 139. Erythropygia leucosticta reichenowi Hartert 1907 : Canhoca, Cuanza Sul. Specimen i imm. $ Chingoroi, gth Sept. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 431 Field Notes This elusive scrub robin was never seen, the single specimen being collected in a net set in slender, very close-growing trees with very little undergrowth, near the stream in the forest. Systematic Notes This is the most southern record for the species and I believe the first known female of this form. It was compared with two males from northern Angola borrowed from Chicago and was found to be a good match, only a trifle paler on the mantle. It still has some young feathers with rufous edges in the secondaries and wing-coverts. 141. Acrocephalus baeticatus cinnamomeus Reichenow 1908. Specimen i $ Chingoroi, 8th Sept. Systematic Notes The three races of Reed Warbler, baeticatus (Vieillot), cinnamomeus and suahelicus Grote, are, at best, only distinguishable in series. This specimen, in moult, is un- satisfactory for comparison but has been identified as cinnamomeus by Mr. C. M. N. White, who is preparing a revision of the species. 143. Bradypterus mariae boultoni Chapin, Ann. Cam. Mus. 31, p. i, 1948 : Mombolo, Cuanza Sul. Specimens 1 c 2 $ Mt. Moco 7,000 feet, i6-23rd Aug. Field Notes This Forest Warbler was quite plentiful in the pockets of forest, as its voice, a loud " chep, chep " frequently repeated, was heard throughout the day, even when other birds were silent. It was not often seen, and always in thick undergrowth and very near the ground. The stomach contained crickets and other insects. 142. Seicercus laurae laurae Boulton 1931 : Mt. Moco. Specimens 2 $ 2 $ i Mt. Moco 7,000 feet, I9~22nd Aug. (i in spirit.) Field Notes Two of these were a pair found in thickish forest, the female carrying nesting material of feathers. Another was one of a party of three or four feeding in foliage about ten feet from the ground. The food was found to be larvae and insect fragments. ZOOL. 6, 7. 26 432 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 145. (a) Calamocichla gracilirostris winterbottomi White 1947 : Manyinga R., Macondo dist., Mexico. (b) Calamocichla gracilirostris cunenensis Hartert 1903 : Cunene R., S. Angola. Specimens (a) 2, 5 m. W. of Munhango, gth Aug. Systematic Notes The taxonomy of the Grey-headed Bush-Shrike has been confused by differences of opinion on the correct use of blanchoti Stephens, poliocephalus (Lichtenstein), or hypopyrrhus (Hartlaub), as the specific name, and on the importance of an orange wash on the underparts, and of a white circle round the eye as taxonomic characters. I follow Grant & Mackworth-Praed (1958 : 19) in using hypopyrrhus as the specific name, and follow Chapin (1954 : 42) in recognizing geographical as well as individual variation in the amount of orange wash on the underparts. I do not however, agree with Chapin 's treatment of the Angola birds, all of which he calls monteiri, distin- guishing them from interpositus of Rhodesia by the lack of any orange wash below on the majority of specimens. The white circle round the eye, characteristic of the type of monteiri (Sharpe), from near Luanda, and the type oiperspicillatus (Reichenow) from Cameroon Mt. he considers an abnormal character of no taxonomic significance. Other specimens with this character are from Dondo and Ndala Tando (in the British Museum), Gabela and " Bucaso " (in Chicago). With the exception of Cameroon Mt. and " Bucaso ", which has not been located, all these places are along or below the escarpment in north-western Angola, and all have riverine and escarp- ment forest rather than woodland. All other specimens examined from Angola in Chicago or London have been taken from above the escarpment in more predomi- nantly woodland areas, and have the white on the face variable and restricted to areas between the eye and the bill, or on the lores. There may thus be an ecological as well as morphological difference between the birds with the white ring and those without and, in my opinion, the name monteiri should be used only for those with this character. The race therefore has a restricted range in north-west Angola and possibly a discontinuous distribution north to Cameroon Mt. (see Serle 1954 : 72). The colour of the underparts of monteiri is variable, the Dondo and Ndala Tando specimens have an orange wash, the others being clear yellow. All other specimens from Angola, without the eye-ring, I would include with interpositus, at least for the present, though noting that the orange wash commonly found in specimens in the east of the range becomes less frequent and less pronounced in specimens in the west. In studying these shrikes I am indebted to Dr. A. L. Rand for a copy of his notes on the species and details of specimens in Chicago. 442 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 225. Neocichla g. gutturalis Bocage 1871 : Huila. Specimens 2 <$ 2 ? 48 m. SW. of Caconda, 28th Aug. (one in spirit). Field Notes For some distance along the Caconda-Quipungo road these Babbler-Starlings were common, flying up from the roadside into bushes and low branches when disturbed by the car. This was in the area of leafless Brachystegia-type woodland mentioned in the discussion on vegetation. Systematic Notes These are the first specimens of the nominate race in the British Museum. In comparison with N. g. angusta Friedmann, they seem to have the head and throat a paler, clearer grey as well as having larger white tips to the tail. 228. (a) Lamprotornis m. mevesii Wahlberg 1857 : Okavango, N. Bechuanaland. (b) Lamprotornis mevesii benguelensis Shelley 1906 : Capangombe, S. Angola. ( L. m. purpureus auct.) Specimens (a) i <$ 25 m. E. of Sa da Bandeira 4,900 feet, 27th Aug.: i $ 10 m. E. of Cahinde 3,000 feet, 6th Sept. (b) i $ 15 m. S. of Quilengues, 3,900 feet, 7th Sept. Systematic and Field Notes The two races of Long-tailed Starling are clearly differentiated in the blue or bronze sheen of the whole plumage and none of the specimens examined show signs of intergradation. The boundary between them seems interesting following the discussion on the three avifaunal zones in Angola, for the bronzy L. m. benguelensis seems to be restricted to what might be termed the southernmost tip of the Escarp- ment Zone, a narrow tongue of rather richer vegetation lying between one thousand and three thousand five hundred feet with the coastal semi-desert on the west and the Brachystegia woodland on the east. Specimens of benguelensis have been obtained at Catengue, near Chingoroi, Quilengues, Vila Arriaga and Capangombe. Quilengues lies more than halfway along the Sa da Bandeira to Chingoroi road which descends gradually with the vegetation becoming progressively richer and more varied northwards. Vila Arriago and Capangombe lie in fertile pockets at the base of the Chela escarpment but bordered westwards by the desert. L. m. mevesii on the other hand has a widespread distribution in the Mopane belt from Nyasaland westwards, and it is confined in Angola to the less arid parts of the Acacia Zone. It has been collected at Jau, Humbe, Chipopia, Mulondo and Mupa. The specimen collected east of Cahinde was one of a number seen on the road down the escarpment, where there was a light covering of mixed bush and stunted THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 443 trees including Mopane. Although Cahinde lies like Capangombe at the foot of the escarpment, and is only forty miles south, it has not the same rich vegetation and the associations of all species collected there are with the Acacia Zone. It should be noted that some of the specimens collected by Monard (1954 : 83, 84) were examined with Captain Grant in 1957 through the kindness of the Director of the Musee d'Histoire Naturelle, Chaux de Fonds. Those from Mupa and Mulondo identified as purpureus were found to be L. m. mevesii, and samples of those identified as mevesii were found to be L. austmlis. 231. (a) Zosterops senegalensis quanzae de Schauensee 1932 : Upper Cuanza R. (b) Zosterops senegalensis heinrichi Meise 1958 : Canzele, Cuanza Norte. (c) Zosterops senegalensis quanzae ^ anderssoni Shelley. Specimens (a) i Mt. Moco 6,500 feet, i5th Aug.: i $ 25 m. N. of Quipeia, 14 Aug. (b) i $ 12 m. S. of Gabela, igth Sept.: i <$ 28 m. N. of Vila Luso, 8th Aug. (c) i <$ Leba, 30th Aug. Systematic Notes Identifications have been made on the basis of Moreau's paper (1957). These specimens have interest in filling some of the gaps that were then noted. The Moco and Quipeio birds have the characteristic dull colouring and large size (wing 65, 62) of typical quanzae. The Gabela and Vila Luso birds are a shade brighter, particularly on the throat and tail-coverts, and the Gabela bird on the forehead. They are close to, but a little more olive than, the Ndala Tando birds quoted by Moreau which have since been named heinrichi. The Leba bird is altogether brighter below, matching in this respect a topotypical anderssoni from Elephant Vley, but being rather darker above and larger (wing 61). 232. Nectarinia kilimensis gadowi Bocage 1893 : Galanga, Benguela dist. Specimens i <$ Vila Flor, I3th Aug.: 9^1 imm. $ i Mt. Moco 6,000-7,000 feet, I5~2ist Aug.: i (J 2 imm. $ i $ i imm. Nova Lisboa, 13-14^ Aug. Field Notes This sunbird was common on Mt. Moco at all levels, feeding on Erythrina and Loranthus. Stomach contents included spiders, Coleoptera, Diptera, and hover-flies. Its call was noted as being exactly like that of the nominate race. 233. Nectarinia bocagei Shelley 1879 : Caconda, Huila dist. Specimens i $ 3 imm. $ near Vila Flor, 13-14111 Aug.: i $ near Luimbale, 25th Aug. 444 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Field Notes Bocage's Sunbird was rare. It was once found feeding on Erythrina with Cinnyris oustaleti and Chalcomitra senegalensis , and once on red Loranthus in Brachystegia woodland. 237. Cinnyris o. oustaleti (Bocage) 1878 : Caconda, Huila dist. Specimens i ? i imm. $ Nova Lisboa, I2th Aug.: i <$ 43 m. W. of Vouga, i2th Aug.: 2 <$ i imm. $ near Vila Flor, I3~26th Aug.: 4 <$ i juv. Mt. Moco 6,000-6,500 feet, i5-24th Aug. i imm. $ near Luimbale, 25th Aug. Field Notes Found feeding on orange-red gladiolis in the Brachystegia woodland on Moco, and on Albizia trees at Luimbale. Its food contained seeds, spiders, beetles and Lepidoptera larvae. Systematic Notes The seven males, including one in which the skull was not fully ossified, are fairly uniform in appearance with a few metallic feathers on the head and mantle and the metallic plumage below confined to the centre of the throat and chest. In four of the specimens sheath feathers of metallic plumage were found on the chest and mantle ; in three, including the immature bird, the wings and tails are in moult, in the others they are fresh. The adult female is moulting wings and mantle. Williams (1955) discusses the moults and plumages of this species, finding that the moult from eclipse to metallic plumage is complete, including wings and tail, but that the wings and tail are not moulted after breeding. It would seem therefore that the males collected are moulting into breeding dress. The finding of these males concurrently with young birds with skulls 10 % or 20 % ossified suggests a protracted breeding season. The identification of the juvenile from Moco is not certain, no comparable speci- mens being available. 246. (a) Anthreptes collaris somereni Chapin 1949 : Anda, Lake Azingo, Gaboon. (b) A.c. phillipsi White 1950 : Lofu R., N. Rhodesia. Specimens (a) i c? i $ 12 m. S. of Gabela, I5~i6th Sept. (b) i $ i $ Luau R., 27-29^1 July ; i imm. $ i imm. $ Lake Dilolo, 5th Aug. i ^ 45 m. W. of Vila Luso, 8th Aug. THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 445 Systematic Notes The adult female from the Luau River has the grey throat and chest characteristic of phillipsi well marked. The Gabela pair have been identified as somereni on locality. In colour they seem indistinguishable from zambesiana (Shelley), though they are smaller. Both are bright clear yellow below, without much olive wash on the flanks, and the female has a pale whitish-yellow throat. Comparison between somereni and zambesiana is a little difficult on available skins in the British Museum since most of those of somereni are small, rather pinched skins which exaggerate the dusky flanks and hide the clear yellow of the tummy ; most of those of zambesiana, and the Gabela pair, are well filled-out skins which show off the yellow. It seems that there may not be such clear-cut differences between the two races as appears from this series. 252. Passer griseus griseus (Vieillot) 1817. Specimen i <$ 12 m. S. of Gabela, zyth Sept. Wing 78, bill 15 mm. Field and Systematic Notes See next Species. 253. Passer diffusus georgicus Reichenow 1904 : Damaraland. Specimen i $ Leba, 3ist Aug. Wing 78, bill 13-5 mm. Field and Systematic Notes The relationship between the two forms of grey-headed sparrow found in coastal Angola has been discussed by Benson (1956) and by White & Moreau (1958), and the authors differ in their views as to whether the two forms can be regarded as conspecific. The two specimens collected conform with Benson's conclusions on the ecological segregation of the forms in western Angola, since the male griseus was obtained in the vicinity of native huts in a clearing in the coffee forest, while the female diffusus was found in a mountain valley some distance from habitation. The Leba bird, like others of diffusus from Angola, noted by White & Moreau, is small and its bill is horn-coloured, paler below, with a yellow gape. The Gabela griseus is also short-winged in comparison with other Angola specimens noted, but has a heavy bill which is black. 257. Ploceus temporalis (Bocage) 1880. Specimens i c? i (with nest), Saiangikilo R., 40 m. W. of Munhango, 9th Aug. Wing 81-84 " bill 19 ; tail 44-45 mm. Bill black ; legs pale reddish-horn ; iris cream. 446 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS Field Notes This rare weaver was only encountered once at a place where the road crossed the river in the centre of an open dambo. The nest, which is unused, was hung in elephant grass beside the river. (Plate 5 B, and 6). Nest. The nest has been examined by Professor and Mrs. N. E. Collias of the University of California, particularly in respect of the possible relationship of P. temporalis with P. capensis. I am grateful to them for the following comments, and for the photograph. " The nest is typically ploceine in its use of green, more or less flexible and inter- woven materials. It is a typical Textor (Ploceus) nest in its ovoidal shape, pensile attachment, ventral entrance, and presence of a special ceiling put in before the nest is lined below. The nest you sent us is unlined and apparently never contained eggs. It belongs to the same size class as nests of Textor cucullatus, T. spekei and T. capensis and not to the size class of nests of the smaller Textor s. " The nest has one specialized feature, unique so far as we know, among nests of the Plocinae. This is the extensive use, for construction of the outer shell, of grass culms from which the flowering tops have apparently been snipped off. " But on the whole the nest is a primitive one for a Textor, in various ways : loose, untidy weaving, use of grass stems in the outer shell, loose fastenings of individual pieces of nest material as well as loose attachments to the support, irregular ceiling of narrow sections of leaf blade, absence of longitudinally-stripped pieces of leaf blade, and in the occasional presence in the ceiling of grass heads (identified as Pennisetum for us by our Botany Department). " In many of these features the nest resembles most closely that of Textor spekei of East Africa. However, the poor ceiling of narrow bits of grass blade seemed to us to be most like that of a young male T. cucullatus. In the presence of grass heads in the ceiling it resembles the nests of Othyphantes, Textor rubiginosa and Textor xanthops ; however, it lacks the abundance of grassheads that is found in the ceiling of these weavers. " The nest of Ploceus temporalis does not seem to us to closely resemble the one nest of Textor capensis that we have been able to examine closely . . . This capensis nest had the outer shell built of grass leaf strips and palm strips rather than grass stems. It had the ceiling made not only of broad (i cm.) grass-leaf sections but also of pine needles and cypress branchlets. We recognize, of course, that a nest is only one clue to the taxonomic affinities of birds." Systematic Notes Having regard to Professor Collias's comments on the differences between the nests of P. capensis and P. temporalis as well as to the differences in the bills of the two forms, I do not feel that temporalis should be regarded as a race of capensis, at least until more is known of its field characters. The unsexed specimen is a good match with the plate in Shelley (1905 : pi. 41) ; both from its size and colour there is little doubt that it is also a male. The other specimen differs in having the olive of the ear-coverts extending on to the chin THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS 447 and centre of the throat making a tongue of dark colour sharply defined from the yellow of the underparts. 258. Ploceus xanthops (Hartlaub) 1862: "Angola" = Lower Cuanza River (see below). Specimens i <$ 2 $ Mt. Moco 6,000-7,000 feet, i5~22nd Aug. 2 $ Leba, 30th Aug.~3rd Sept. i $ Chingoroi, 8th Sept. Systematic Notes The type locality of Ploceus xanthops is given only as " Angola." The original description however is among notes on a small collection made by Monteiro of birds from the coastal areas of north Angola. Definite localities are given for other species, either Cuanza River, Cambambe or Massangano, while the weaver is noted as being " common everywhere . . . particularly over water." It seems that the type locality can therefore be usefully restricted to the Lower Cuanza River and specimens collected from Luanda, Dondo and Vila Salazar can be considered as typical. This is important since examination of skins in the British Museum, and some from Chicago, show that there is considerable geographical variation within Angola, particularly in the females. The variation is particularly interesting since it follows exactly that found in the Zosterops of Angola. Specimens of typical xanthops in the north are rather a yellow green matching the Zosterops described by Moreau (1957 : 372) as intermediates between Z. s. kasaica and Z. s. quanzae and since named as heinrichi Meise (1958 : 80). Those of the Benguela highlands are a darker and duller green, matching Z. s. quanzae, while those of the south are brighter, more washed with yellow as is Z. s. anderssoni. When specimens of Ploceus xanthops from the rest of Africa were examined, including some borrowed from Pretoria, it was found that this variation in Angola represented the extremes of variation in colour found throughout the continent. The populations of north-eastern Africa, which have sometimes been separated as cambumi (Sharpe), are brighter and more yellow than typical xanthops but similar to the southern birds. The populations of southern Africa, jamesoni (Sharpe), are very slightly greener and duller than typical xanthops with usually slightly less suggestion of dark streaks on the mantle but not so green or dull as the highland birds of Angola. Most specimens from central Africa match typical xanthops, but occasional birds, nearly all of which are from highlands of either southern Tangan- yika or Nyasaland, are dull and green, very close to the highland birds of Angola. This is again consistent with Moreau's findings on dark populations of Zosterops at higher altitudes (loc. cit.: 370). In view of the sporadic distribution of darker and brighter populations it is difficult to see how any races can be recognized on colour differences. In size the wing measurements of all males examined are between 86 and 96 mm. and all females ZOOL. 6, 7. 27 448 THE ECOLOGY AND TAXONOMY OF SOME ANGOLA BIRDS between 80 and 92 mm. with the exception of three from Maun and Lake Ngami, the male of which has a wing of 102 and the females 92 and 94. The name maunensis (Roberts), was given to birds from Maun and from the Chobe River ; the latter are no larger than typical xanthops so that if maunensis is to be considered distinct from xanthops it must be restricted to Lake Ngami. In my opinion it is hardly consistent in an admittedly variable species to single out the characters of this one population for recognition by name. The following table shows that there are other local variations in size but none of any significance. An attempt has also been made in the table to indicate the colour variation in the populations though it must be appreciated that there is also con- siderable individual variation particularly in the amount of yellow on the forehead of males. TABLE. Populations of Ploceus xanthops Wing length A. ( \ Area 6* ? (1) Angola, north . 12 $ 7 $ . 86-93 80-85 (2) highlands . 3 6* 5 $ 90-91 82-89 (3) south . 10 ' Tj* 14 54 E. Cuanza Norte. /^\ 1 " t *' k 3A ^ g ] \ 2> PLATE 5 A. Escarpment at Leba. Habitat of Xenocopsychus ansorgei. B. Dambo 40 miles west of Munhango, with nests of Ploceus temporalis. PLATE 6 Nest of Ploceus temporalis (Photo. N. E. Collias) Bull. B.M. (N.H.) Zool. 6, 7 PLATE 5. Bull. B.M. (N.H.) Zool. 6, 7. PLATE 6. I (Care^^S^/l ^Wmfe* <^iS M-%rov ^rS :&# *m?w&7. && INDEX TO VOL: VI The page numbers of the principal references and the new taxonomic names are printed in CLARENDON type. NOTE. The paper by H. E. Quick on British Slugs is indexed on pp. 224-226. abingoni, Campethra \ . abyssinica, Hirundo abyssinicus, Pseudoalcippe Accipiter Achaetops \ . aciculatus, Sejus Acrocephalus . acuticaudus, Lamprotornis adustus, Alseonax aereus, Ceuthmochares . aethiopicus, Laniarius afer, Cinnyris afer, Nilaus . afer, Parus afer, Pternistes . . 386, afra, Pytelia . africana, Mirafra . . 386 africanus, Bubo africanus, Rhinoptilus Agapornis Alaudidae albicauda, Elminia albigularis, Hirundo albilinea, Columba fasciata albinucha, Columba albitorques, Columba albiventris, Halcyon albofasciata, Certhilauda albogularis, Francolinus 386 albogularis, Poliospiza Albula . 329, 332, 334- Albuloidae Alcedo .... Alepocephaloidae . . 331 Alepocephalus Alethe alosoides, Hiodon . Alseonax . . 380, 384, alstoni, Sejus altivelis, Plecoglossus ameliae, Macronyx amethystina, Chalcomitra Amia .... Anaspides anchieti, Anthreptes anchietae, Buccanodon . Anchoviella . 382, 388 384, 402, 436 382, 392 . 382 388, 395 . 50, 76-78 386, 431 - 384 . 380 . 382 380, 400, 438-439 382, 384,392 380, 390,399 ,\ . 390, 400 388, 398, 399, 404 . 388 394, 399, 418-419 . . .388 . 388 . 388 44 380, 384, 390 . 386 20 5, 13, I* 3, 8, 11 . 386 386, 393-394 393-394- 4 5-409 390 335- 357- 359-36o 357- 360 406 335, 357-358,360 330-332, 335, 337 . 384 333, 341 388, 404, 405, 423 51,74-76 . 340 386, 393-394- 421 . 382 328, 343 . 318 - 382 380, 414 335-336 Andropadus . angolensis, Dryoscopus . angolensis, Gypohierax . angolensis, Hirundo angolensis, Mirafra . 386, angolensis, Monticola anguitimens, Eurocephalus annandalei, Indomysis anophelinae, Antromysis Anotopteridae ansorgei, Sylvietta ansorgei, Xenocopsychus Anthreptes . Anthus Antichromus antillanus, Sejus Antromysis Apalis . . . 384, Apaloderma . Aplochiton . . 333, 339, Aplochitonidae Aplopelia araucana, Columba ardens, Coliuspasser Argentinoidae 331, 338, 345, are o la t us, Sejus argenteus, Photichthys . argentina, Columba Argentina Argentinidae aridula, Cisticola . Arionidae, Index to armiger, Petalophthalmus arnotti, Myrmecocichla . arnotti, Thamnolea arquatrix, Columba astrild, Estrilda Astronesthoidae atricollis, Eremomela atricollis, Ortygospiza atrococcineus, Laniarius atrogularis, Serinus Aulopus auratus, Oriolus aurea, Euplectes australis, Hyliota . australis, Lamprotornis . 382, 392, 405 384 406 . 386 393-394. 409, 419 380, 390, 400 377. 38o 3I3-3M 3 357 . 433 377, 388, 395. 429 382,384, 444-445 386,404, 420-421 . 380 . 51, 78-8o 299, 300 388, 392, 404, 432 382, 412-413 34 1 . 347-349, 35 346-351 22 . 6, 20 . 386 352, 356, 358, 360 . 51, 64-65 . 337-338 . 5. 15 338-339 . 345-346 386, 390 224-226 311-312 377. 38o 377 5. 13, 14 388, 406, 449 . 358, 360 380, 390, 434 . 388 390 . 388 344 . 382 390, 395. 448 380, 386, 428-424 443 456 australis, Tchagra . axillaris, Coliuspasser ayresii, Cisticola INDEX baeticatus, Acrocephalus barbata, Erythropygia . barbatus, Pycnonotus Bathyclupeidae . . Bathylaconidae Bathynellacea Batis . . 380, 384, 391 bellingeri, Sejus bengalus, Uraeginthus . bicinctus, Eremialector . bicirrhosum, Osteoglossum bicolor, Ducula bicolor, Laniarius . 390 bicolor, Symplectes bifasciatus, Cinnyris bocagei, Cossypha bocagei, Nectarinia bohmi, Myopornis . bohmi, Neafrapus . bollii, Columba borealis, Sejus Boreomysinae . 299 Boreomysis . bottazzii, Spelaeomysis . brachyptera, Cisticola Bradornis Bradypterus . brevicaudata, Camaroptera brevipes, Monticola brevirostris, Schoenicola bronni, Leptolepsis browning!, Sejus brunneus, Dioptrornis Bubo .... Buccanodon bullions, Cisticola . burtoni, Poliospiza 390 386 386 . 386, 431 . 380 382, 400 357 358 319 392, 397. 424-425 5L 81 . 388 . 388 . 341-343 15 400, 404, 438-439 384. 39i 384 384, 391-392 380, 390, 443-444 . 380 406 4, 8, 9 50, 56 , 304-305, 37-38 . 310-312 299, 308 . 386 380, 388 384, 392, 431 380, 384, 388, 397, 399, 434-436 388, 400 . 386 359 . 50, 86-88 384, 392 380, 414 386 384. 392 caballus, Hormyrus cabanisi, Emberiza cafer, Cuculus caffer, Micropus . cailliautii, Campethera . Calamocichla Calamonastes Calandnella . calmani, Micraspides caluus, Gymnobucco Camaroptera 380, 384, Campephaga . , Campethera . capensis, Euplectes capensis, Fringillaria capensis, Motacilla 342-343 . 382 . 380 .386 380, 416-417 . 386, 432 380, 388, 432 386, 404 302-303 382, 391, 414-415 388, 397- 399, 434-436 . 384, 405 380, 382, 388, 416-417 . 386 390,45 . 386 capensis, Ploceus . capensis, Smithornis capillatus, Platyseius Caprimulgus . 380, 382, 386 caribaea, Columba cassini, Alseonax . cassiteridum, Sejus castanotus, Colius 380, 382, cayennensis, Columba See also rufina, Columba cenotensis, Antromysis . Centropus Cercomela Certhilauda . Ceuthmochares Chalcomitra . chalybeus, Cinnyris Chanoidae Chanos Cheiroseius . chelicuti, Halcyon . chiniana, Cisticola . Chirocentrus . Chlorocichla . 384, 391, 392, Chloropeta Chlorophoneus . 380, 384 Chrysococcyx chrysoconus, Pogonolius Ciccaba Cichladusa cincta, Riparia . cinerea, Apalis cinerea, Calandrella cinereovinacea, Clytospiza cinereus, Alseonax cinereus, Haplochromis . Cinnyricinctus Cinnyris . 382, 384, 390 clayi, Sejus . Cisticola 380, 384, 386, 390 Clupea .... Clupei .... Clupeidae Clupeiformes Clupeoidae Clupeoidei Clytospiza Coccopygia . collaris, Lanius coqui, Francolinus Coilia .... collaris, Anthreptes Colius . . 380, 382, 388, Coliuspasser . Columba concolor, Corythaixoides concreta, Dyaphorophyia Coracias Coracina Coregonidae . . 446 382, 391 34 392, 405, 411-412 . 6, 20 . 384 . 51, 61-62 388, 391, 393, 402, 413 20 . 299 . 386 380, 388, 397, 399 386, 393-394 . 382 . 382 . 382 336-337, 357, 360 336-337, 353-354 28 . 380 380, 390, 436 335-336 393, 396-397, 422 384 392, 406, 439-440 . 380 380, 388 . 382, 410 . 384 . 386 . 384, 392 386, 404 388, 404, 449 388, 404, 405, 423 229-230, 239-242 382 392, 401, 406, 444 30, 31, 50, 73-74 39i, 395, 44> 436 335-336 357-358, 359-36o 328, 345 327, 356-365 335-336, 357, 360 331-333, 357, 360 388, 404, 406, 449 . 388 386, 397, 399 . 380 333 382, 384, 444-445 39i, 393, 42, 413 386, 393-395, 44 . 1-23 388 384, 392, 425 . 380 . 380 346 INDEX Coregonus corensis, Columba . coryphaea, Viridibucco . corythaix, Tauraco Corythaixoides Corytheola Cossypha . . 384, 391- cotti, Heteromysis cristata, Corytheola croceus, Macronyx Cromeria Cromerioidae cruzi, Francolinus See swierstrai, Francolinus Cryptospiza . cubla, Dryoscopus . cucullata, Lonchura cucullatus, Ploceus Cuculus cupreus, Cinnyris . Cursorius curtipes, Sejus Cyanomitra . Dallia . Dallioidae Dalliidae dambo, Cisticola . decemremiger, Zerconopsis delagorguei, Columba Dendropicos . diadematum, Tricholaema Dialiptila Dicrocercus . Dicrurus diffusus, Passer dimidiata, Hirundo Dioptrornis . dorab, Chirocentrus dromadis, Sejus Dryoscopus . Ducula Dussumieria . Dyaphorophyia elegans, Erythrops Elminia Elopoidae Elopoidei . . 331, Elops . . 328, 330-332, elphinstonii, Columba Emberiza emini, Cisticola empodisima, Haplochromis Epinephelus Episeiella Episeius Eremialector Eremomela . 380, 388, 390, 338 . 6, 19 382, 392, 415 382, 391, 409 . 388 406 392, 405, 429-430 299-300 406 393, 394 333. 354 357-360 384, 393, 448-449 380, 384, 390 . 382 390 . 380 - 382 . 386 . 52, 59-6i 384, 406 333 352, 358 331. 333 . 386 . 91, 96-98 7, 22 380, 388, 417 380 ' 3 . 380 384 390, 400, 445 . 386 384. 392 335-336 . 50, 62-64 380, 384 i5 336-337. 353 384, 392, 425 320 380, 384, 390 357- 360 334-335, 357. 360 334, 356, 358-359 . 5, 14 382, 390 . 386 262-266 343-344 28 28 . 388 399, 405, 433-434 Eremopterix . Erythrocnus . erythrolophus, Tauraco 382, Erythrops erythrops, Cisticola Erythropygia 380, 384, 388, Esocoidae Esocoidei . 331-333, 341, Esox . . . 330-333, Estrilda Eucarida Eucopia Eucopiidae Eumalacostraca Euphausia Euplectes . . 386, 390, Eupodotis Eurocephalus eversmanni, Columba 457 386, 388, 395 406 392-393, 409-4io 3 20 . 384, 391 398, 405, 430-431 352, 358 349-352, 358, 360 341, 349, 358-359 384, 388, 406, 449 . 318 311, 313-314, 319 299, 305, 37. 313 318-319 319 395, 44, 4 6 , 44 8 . 406 377. 38o 4, 8, 10 falkensteini, Chlorocichla 384. 392-293, 396-397, 422 familiaris, Cercomela . 380, 388, 397, 399 fasciata, Columba . . . .6, 19, 20 fasciatus, Gallaxias .... 340 fasciolatus, Calamonastes . . . 388 finschi, Francolinus . . 380, 390, 408 fischeri, Phyllastrephus . . . 384, 391 flavicans, Prinia . . . 380, 390, 399 flavicollis, Chlorocichla . . . 384, 391 flavida, Apalis .... 388, 404 flavigaster, Hyliota . . . 380, 386 flavirostris, Columba . . . . 7, 20 flaviventris, Chlorocichla . . . 384 naviventris, Emberiza . . . 382, 390 flavivertex, Serinus .... 388 fossii, Caprimulgus . . . 386, 411 Francolinus 380, 382, 386, 388, 390, 392, 393-394, 45, 407-409 Fringillaria .... 390, 404, 450 Fringillidae . . . . . .402 fulleborni, Macronyx . . 386, 393-394 fuligula, Ptynoprogne . . . 386, 436 fulvescens, Malacocincla . 382, 392, 422 fulvicapilla, Cisticola . . . 380, 390 fulviventris, Phyllastrephus 384, 392-393. 396-397, 423 fuscescens, Dendropicos . . 380, 388, 417 fuscus, Cinnyris ..... 390 fyllae, Hansenomysis . . . 311, 314 gabela, Muscicapa . . . 396-397 gabela, Prionops 378, 384, 390, 396-397, 400, 406, 437-438 Galaxias . 333, 339-341, 345, 347-35 1 Galaxiidae .... 345-35O, 357 Geokichla .... 380, 390, 428 45 INI )EX Glareola 386 icteropygialis, Eremomela 388,399,405,438-434 Glaucidium ...... 380 impetuana, Fringillaria .... 390 Gnathophausia . . . 305, 307, 317 Indicator . . . 382, 382, 415~4i6 Gonorhynchoidae . . . 337, 357, 360 indicus, Colius ..... 388 Gonorhynchoidei . 331-334, 352-354, 357, 360 Indomysis ..... 313-314 Gonorhynchus . . . 337, 352-356 infuscata, Bradornis .... 388 Gonostoma ...... 337 ingens, Gnathophausia . . . -317 Gonostomoidae .... 358, 360 inornata, Columba . . . . 7, 20 goodsoni, Columba . . . . 7, 21 insignis, Prodotiscus . . . 380, 398 gracilirostris, Calamocichla . . 386, 432 iriditorques, Columba . . . . 7, 21 granatina, Granatina .... 390 ishmaeli, Haplochromis . 230, 273, 275-279 grayi, Salvelinus ..... 331 Isospondyli ..... 827-365 greyi, Gonorhynchus . . 337, 352-353 Ispidina ...... 382 grimwoodi, Macronyx italicus, Ameroseius .... 42 386, 393-394. 405. 421-422 italicus, Plesiosejus . 28, 29, 31, 42, 43-45 griseiventris, Parus .... 400 griseocephalus, Mesopicos . 380, 382, 418 griseoflava, Eremomela .... 434 jamaicensis, Platyseius . . -34, 89-41 griseopyga, Hirundo .... 386 jamaicensis, Sejus . . . .51, 88-84 griseus, Passer . . . 384, 406, 445 jamesoni, Lagonosticta .... 388 guinea, Columba . . . . 3, 4, 8, 11 j_l /* 1 i_ /-An janthma, Columba . . . 3, 6, 16 guinea, Dialiptila ..... 3 gularis, Poliospiza .... 382, 390 gutturalis, Neocichla ' . . 377, 382, 442 jardinei, Turdoides . . . 377, 380 Jenkinsiella . . . . . -336 jouyi, Columba . . . . . 6, 18 Gymnobucco . . 382, 391, 414-415 junioniae, Columba . . . . 5, 12 Gypohierax ...... 406 Halcyon . . . . . 380, 386 Karrucincla ..... 388, 395 halmaheira, Columba vitiensis . 16, 18 kathleenae, Batis .... 424-425 Hansenomysis . 307, 311, 313-314, 316 Haplochitonidae .... 34S-34& Haplochromis .... 229-281 kennedyi, Sejus . . . -51, 65-66 kilimensis, Nectarinia . . 382, 390, 443 klaasi, Chrysococcyx .... 380 Haplomi Kneria 337, 353~354 See Esocoidei Knerioidae . . . 886-337, 357, 360 Harengula . . 335~336, 345 krameri, Umbra ..... 341 hartlaubi, Coliuspasser .... 386 hartlaubii, Pteronetta . . . . 406 Heteromysis .... 299, 304 labradorensis, Zerconopsis . 31, 91, 98-ioo Heterotis , 332, 341-343, 359 Labrochromis . . . . .275 Heterotrogon . . . 382, 392, 412 lacrimosus, Haplochromis . . 280-233 heuglini, Cossypha . . 384, 391, 405 laelaptoides, Sejus . . .32, 51, 54-55 Hiodon . . 330-333, 341, 353, 358-359 Lagonosticta . . 388, 390, 400, 406 Hiodontoidae . . . 341, 358-860 lais, Cisticola ..... 386 hirundineus, Dicrocercus . . . 380 Lampanyctus ..... 344 Hirundo . . 384, 386, 402, 404, 436 Lamprotornis . 382, 384, 390, 399, 442 hodgsonii, Columba . . . 5, 13, 14 landanae, Lagonosticta . . . 388, 400 holthuisi, Rhopalonurus . . . . 287 Laniarius . 380, 390, 399, 400, 404, 488-439 Stygiomysis Laniidae ...... 402 286-287, 299, 303-305, 310-313, 319- Lanioturdus .... 390, 395 320, 234, PL 3 Lanius . 380, 384, 386, 390, 392, 397, 399 horridus, Plesioejus . . 42, 48-49 Lasioseius ...... 34 hulli, Sejus . . . . .50, 81-82 lateralis, Cisticola ..... 386 humilior, Haplochromis . . . 248-252 latirostris, Andropadus . . . 382, 392 hydruntina, Stygiomysis 286, 297-299, 301, 304 laurae, Seicercus . 384, 391-392, 402, 431 Hyliota .... 380, 428-424 Lepidopidae . . 299-300, 304-808-309, 313 Hypargos ...... 384 Lepidops . . 299-309, 312-313, 316 Hyphanturgus .... 384, 402 lepidops, Spelaeogriphus 321-322, 324, PI. 4 Hypochaera ..... 384, 406 Lepidopthalmidae Hypomesus ..... 334, 339 See Lepidopidae hypopyrrhus, Malaconotus 380, 397, 399, 441 Leptolepsis ...... 359 leucocephala, Columba . leucogaster, Cinnyricinchus leucomelas, Parus . leucomelas, Tricholaema leuconota, Columba leucophrys, Erythropygia leucopogon, Prinia leucopygia, Turdoides Leucosoma leucosticta, Erythropygia Leuroglossus levaillantoides, Francolinus libonyanus, Turdus Limacidae, Index to lineiventris, Anthus Liraspis litsipsirupa, Geokichla livia, Columba Lonchura longicauda, Elminia longuemariae, Anthreptes Lophogaster Lophogastrida Lophogastridae Lovettia 332, 339, 341, 345- lucius, Esox . ludwigii, Dicrurus . luhderi, Laniarius . Lybius .... 3- 6, 18 . 382 380. 397- 400 388, 395 3,4,8,11 388, 398, 405, 430 . 384 386, 388 339 384, 430-431 339 . 386 . 380, 400 224-226 380, 421 27 380, 390, 428 3,4,8,11 . 382 . 400 . 382 307, 3i3 299, 304-309 . 299 350, 351, 356, 360 . 332, 341 . 384 . 399 380, 418-441 mackinnoi, Lanius . . . 384, 392 Macristiidae ...... 358 Macristium ...... 356 Macronyx . 386, 393-394, 404, 405, 421-422 Macropinna . . . . . -338 macrops, Haplochromis . . 286-239, 242 Macrosphenus .... 396-397 maculosa, Columba . . . 3, 6, 19 maderensis, Columba palumbus . . 9 mahagiensis, Haplochromis . . .273 mahali, Plocepasser . . . 390, 405 major, Plesiosejus . . . 42, 46, 47-48 Malacocincla . . . 382, 392, 422 Malaconotus . . . 380, 397, 399, 441 malacophagos, Haplochromis . . .269 malherbii, Columba . . . . 7, 21 margaritae, Batis 384, 391-392, 397, 424-425 mariae, Bradypterus . . 384, 392, 431 martini, Haplochromis . . . 229, 245 mayeri, Columba . . . . . 7, 22 Megalops . . . 334~335 Melaenornis ...... 380 melampyra, Tchitrea 384, 397, 400, 426-427 melanotis, Coccopygia . . . .388 melba, Pytelia .... 382, 390 Melittophagus ..... 380 Melocichla ...... 386 mentalis, Melocichla .... 386 Merops ,,,,.. 380 INDEX 459 Mesopicos .... 380, 382, 418 mevesii, Lamprotornis 382, 390, 399, 442-443 meyeri, Poicephalus . . 380, 382, 405 michaeli, Haplochromis .... 262 michaeli, Zerconopsis . . .92, 95-g6 Micraspides ..... 302-303 microdon, Salangichthys . . . 351 Micropus . . . . . .386 Microstoma ...... 345 Microstomidae .... 346-346 minor, Indicator . . .382,388,415-416 minulla, Batis .... 384, 397 minullus, Accipiter .... 382 minutus, Antichromus .... 380 mirabilis, Thermosbaena . . 303, 318 Mirafra 386, 388, 393-394, 399, 405, 409, 418-419 mitrata, Numida ..... 388 molitor, Batis .... 380, 397 mollicomus, Platyseius . . . 34, 37-39 monteira, Clytospiza .... 406 monteiri, Tockus .... 388, 492 Monticola . . . 380, 388, 390, 400 monticola, Oenanthe .... 388 Mormyriform.es Mormyrus Motacilla . . motitensis, Passer . mozambicus, Serinus muestairi, Zerconopsis multicolor, Chlorophoneus Muscacapidae Muscicapa .... Myopornis .... Myrmecocichla 377, 380, 386, 393 332, 384, 304, 285-286, 300-301, . 299-300, 304- Mysida Mysidacea Mysidae Mysidella Mysidetes Mysidopsis Mysis . nabouroup, Onychognathus Nansenia narina, Apaloderma natalensis, Caprimulgus . natalensis, Chloropeta natalensis, Cisticola natalensis, Cossypha Neafrapus Nebalia neborealis, Sejus necorniger, Sejus Nectarinia Nectariniidae Nematolosa . Neocichla Neomysis 333, 343, 357 342-344, 353 386 390, 405 . 388 31, 91, 94-95 392, 439-440 402 396-397 380 -395, 428-429 307-309, 313 304, 306, 318 35, 37- 3*3 34 304, 37 300, 307, 316 37 390, 395 338 382, 412-413 386, 405, 411 384 . 386 384, 391, 430 406 . 319 .50, 84-85 29, 52, 55-56 382, 390, 443-444 402 335 377, 382, 442 300 460 Neoscopelus . nepalensis, Sejus . Nesoenas . . Nicator niger, Parus . nigricans, Pycnonotus Nilaus .... nitidula, Lagonosticta nigerrima, Hypochaera . nigra, Myrmecocichla nigricans, Haplochromis nigricollis, Hyphanturgus nigrifons, Chlorophoneus nigrirostris, Columba nigrorufa, Hirundo niloticus, Haplochromis . niloticus, Heterotis nitens, Columba janthina nitens, Lamprotornis niveoguttatus, Hypargos nivosa, Campethera nodosus, Sejus norfolciensis, Columba Notopteridae Notopteroidae Notopterus .... Novumbra .... nubicoides, Merops nuchisquamulatus, Haplochromis Numida obtusidens, Haplochromis ocularis, Hyphanturgus . Oenanthe oenas, Columba Oenoenas oenops, Columba . olidus, Hypomesus olivacea, Cyanomitra olivaceus, Turdus . oliviae, Columba Onychognathus Opisthonema Opisthoproctoidae See Argentinoidae Opisthoproctus orientalis, Psalidoprocne Oriolus orix, Euplectes o mat us, Sejus Ortygospiza . Osmeridae Osmeroidae Osmerus Osteoglossi . Osteoglossoidae Osteoglossoidei Osteoglossum Otus . oustaleti, Cinnyris INDEX 341- 344 . 51, 71-72 - 3, 7, 22 384 390, 400 388, 400 380, 390, 399 . 388 384, 406 . 386 245 . 384 384, 406, 440 . 7,21 386 243-245 342 16 390 384 . 382, 417 . 51, 85-86 3. 6, 16 356 342, 358-860 342 333 . 380 245 388 265, 266-269 384, 402 386, 388 4, 8, 10 . 3, 7, 20-21 . 7, 20 . 334 384 . 384, 400 4, 8, 10 390, 395 335 338 . 380 382 386, 406 31, 50, 88-89 345-346, 350 331, 339-341, 349-352, 358, 360 345 358, 359, 360 341-343, 358-360 331-333, 358, 359 332, 341-343, 353 . 388, 410 . 382, 444 paena, Erythropygia pallidiceps, Columba pallidirostris, Tockus pallidus, Bradornis pallidus, Haplochromis . paludicola, Estrilda palumboides, Columba . palumbus, Columba pammelaina, Melaenornis Pantodon paradisaea, Steganura . 388, 405 2, 6, 18 . 380 380, 388 233-236, 252, 255 . 388 . 5, 15 4, 8, 7 . 380 34 I ~343> 356 382 parbatensis, Sejus . . . -51, 72-73 Parisoma Parus . Passer . Passeriformes passerina, Mirafra . pectoralis, Caprimulgus pectoralis, Coracina pelios, Turdus . peltrata, Platysteira Peracarida perlatum, Glaucidium perreini, Estrilda . Petalophthalmidae Petalophthalmu s Petrochelidon Petronia 384, 388 380, 382, 390, 397, 400 384, 400, 405, 445 402-405 388, 405 . 380 . 380 384, 391, 400, 427-428 . 380, 384 318-321 . 380 384 299, 304-305, 307-308, 316 307, 311-313 386, 404, 436-437 . 382 . 50, 66-68 230, 270-275 329, phalangioides, Sejus pharyngomylus, Haplochromis phoenicea, Campephaga .... 384 Photichthys 337-338 Phractolaemoidae . . 386-337, 357, 360 Phractolaemus . . 333, 337, 353-354 Phyllastrephus 384, 391, 392-393. 396-397, 423 picazuro, Columba . . . . 6, 19 picta, Ispidina . . . . . 382 pileata, Oenanthe ..... 386 pipiens, Cisticola ..... 386 placodus, Haplochromis . . . 273-274 Platyseiinae ..... 27-88-101 Platyseius .... 30-31, 34-42 Platysteira ..... 380, 384 Plecoglossidae .... 346, 350 Plecoglossus ..... 339, 346 Plesiosejus .... 30-31, 34, 42-49 Plocepasser ..... 390, 405 Ploceus . . 382, 384, 386, 390, 445-448 plumata, Prionops .... 380 plumbea, Columba . . . . 7, 21 plumbeum, Parisoma .... 384 Pogonolius ..... 380, 388 Poicephalus .... 380, 382, 405 poliocephala, Alethe . . . .384 poliocephalus, Caprimulgus . 382, 392, 411 polioptera, Cossypha . . 384, 429-43O Poliospiza . 382, 384, 390, 392 pollenii, Columba . . . . 5, 13, 14 pratincola, Glareola .... 386 preussi, Petrochelidon .... 437 Prjnia , , . 380, 384, 390, 399, 405 INDEX 461 pririt, Batis . Prodotiscus progne, Coliuspasser Prototroctes Psalidoprocne Pseudoalcippe Pternistes Pteronetta Pterothrissus Ptynoprogne pulchricollis, Columba pulitzeri, Macrosphenus punicea, Columba . pusillus, Melittophagus Pycnonotidae Pycnonotus . pycnopygius, Achaetops Pytelia quadribachys, Alcedo quinterensis, Typhlolepidomysis quiscalina, Campephaga regulus, Prodotiscus reichenovii, Cryptospiza . remiger, Gamasus . remiger, Zerconopsis Retropinna . . 333, Retropinnatidae Retropinnidae retzii, Prionops Rhinoptilus . . * Rhopalonurus Rhopalophthalmu s Riparia riponianus, Haplochromis robusta, Cisticola . Rosauridae . roseicollis, Agapornis rostrata, Boreomysis rostratus, Alepocephalus rubricata, Lagonosticta rufescens, Sylvietta ruficapilla, Spermophaga ruficapilla, Sylvietta ruficauda, Cichladusa ruficinnamomea, Mirafra rufigula, Petrochetidon rufilata, Cisticola . rufina, Columba See also cayennensis, Columba rufiventer, Erythrocnus . rufiventris, Parus . rufogularis, Apalis . . rupestris, Columba